# ![@akshay_pachaar Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::703601972.png) @akshay_pachaar Akshay 🚀

Akshay 🚀 posts on X about ai, microsoft, data, science the most. They currently have [-------] followers and [---] posts still getting attention that total [-------] engagements in the last [--] hours.

### Engagements: [-------] [#](/creator/twitter::703601972/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::703601972/c:line/m:interactions.svg)

- [--] Week [-------] +41%
- [--] Month [---------] -15%
- [--] Months [----------] -16%
- [--] Year [----------] +53%

### Mentions: [--] [#](/creator/twitter::703601972/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::703601972/c:line/m:posts_active.svg)

- [--] Month [--] -39%
- [--] Months [---] +13%
- [--] Year [---] +124%

### Followers: [-------] [#](/creator/twitter::703601972/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::703601972/c:line/m:followers.svg)

- [--] Week [-------] +0.50%
- [--] Month [-------] +1.50%
- [--] Months [-------] +11%
- [--] Year [-------] +35%

### CreatorRank: [------] [#](/creator/twitter::703601972/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::703601972/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[technology brands](/list/technology-brands)  #5553 [stocks](/list/stocks)  16.53% [social networks](/list/social-networks)  8.26% [finance](/list/finance)  4.13%

**Social topic influence**
[ai](/topic/ai) 28.93%, [microsoft](/topic/microsoft) 11.57%, [data](/topic/data) 11.57%, [science](/topic/science) 9.92%, [llm](/topic/llm) #71, [build a](/topic/build-a) 4.96%, [if you](/topic/if-you) 4.96%, [agentic](/topic/agentic) #320, [$googl](/topic/$googl) 4.13%, [education](/topic/education) 4.13%

**Top accounts mentioned or mentioned by**
[@akshaypachaar](/creator/undefined) [@mlspring](/creator/undefined) [@everlier](/creator/undefined) [@mr_cryptox](/creator/undefined) [@andreifedyna](/creator/undefined) [@metekkizilkaya](/creator/undefined) [@osamara6991644](/creator/undefined) [@royshilkrot](/creator/undefined) [@3blue1brown](/creator/undefined) [@chipro](/creator/undefined) [@avichawla](/creator/undefined) [@tigerdatabase](/creator/undefined) [@tricalt](/creator/undefined) [@bwcdeals](/creator/undefined) [@roopaknijhara](/creator/undefined) [@pashxme](/creator/undefined) [@shail_2302_](/creator/undefined) [@thiagotm](/creator/undefined) [@clawrecipes](/creator/undefined) [@avery_pelle](/creator/undefined)

**Top assets mentioned**
[Microsoft Corp. (MSFT)](/topic/microsoft) [Alphabet Inc Class A (GOOGL)](/topic/$googl)
### Top Social Posts
Top posts by engagements in the last [--] hours

"A 100% open-source alternative to n8n Sim is a drag-and-drop UI for creating powerful AI agent workflows: - Runs locally on your machine - Works with local LLMs I built a stock market research agent & connected it to Telegram in minutes. Here's a step-by-step guide:"  
[X Link](https://x.com/akshay_pachaar/status/2000569778260578538)  2025-12-15T14:12Z 247.1K followers, 179.1K engagements


"http://x.com/i/article/2012485864727977984 http://x.com/i/article/2012485864727977984"  
[X Link](https://x.com/akshay_pachaar/status/2012514342730904036)  2026-01-17T13:16Z 246.8K followers, 96.3K engagements


"clone any voice with a 5-second audio clip. VoxCPM is an open-source project that takes a fundamentally different approach to text-to-speech. most TTS systems convert speech into discrete tokens. this creates a bottleneck that limits how natural the output can sound. VoxCPM skips tokenization entirely. it models audio in continuous space using an end-to-end diffusion autoregressive architecture. the result is speech that actually sounds human. here's what makes it special: context-aware generation: it reads your text and infers the right prosody emotion and pacing automatically. no manual"  
[X Link](https://x.com/akshay_pachaar/status/2012865076420923518)  2026-01-18T12:30Z 246.7K followers, 81.7K engagements


"this is huge. ollama is now compatible with the anthropic messages API. which means you can use claude code with open-source models. think about that for a second. the entire claude harness: - the agentic loops - the tool use - the coding workflows all powered by private LLMs running on your own machine. https://twitter.com/i/web/status/2013311949770215512 https://twitter.com/i/web/status/2013311949770215512"  
[X Link](https://x.com/akshay_pachaar/status/2013311949770215512)  2026-01-19T18:05Z 246.4K followers, 171.7K engagements


"If anyone is curious the rationale behind diffusion LMs is that autoregressive models can't backtrack. One bad token at position [--] and everything after snowballs from that mistake. Diffusion LMs start with noise and refine the entire sequence simultaneously across multiple passes so inconsistencies can be corrected before finalization. https://twitter.com/i/web/status/2016405229403832611 https://twitter.com/i/web/status/2016405229403832611"  
[X Link](https://x.com/akshay_pachaar/status/2016405229403832611)  2026-01-28T06:57Z 246.5K followers, [----] engagements


"Build agents that can actually do real-world tasks Agent Reinforcement Trainer (ART) is a framework to train multi-step LLM agents for real-world tasks using GRPO. Just a few lines of code. No manual rewards needed. vLLM + Unsloth combined 🚀 100% open-source"  
[X Link](https://x.com/akshay_pachaar/status/2016500285599011126)  2026-01-28T13:15Z 247.5K followers, 37.6K engagements


"DevOps vs. MLOps vs. LLMOps: Many teams are trying to apply DevOps practices to LLM apps. But DevOps MLOps and LLMOps solve fundamentally different problems. Here's why this matters: 88% of ML initiatives struggled to reach production using traditional DevOps practices. And LLMs introduce challenges that even MLOps wasn't designed for. Let's break it down: DevOps is software-centric. You write code test it and deploy it. The feedback loop is straightforward: Does the code work or not The primary artifact is code. Testing is deterministic. The tooling is mature after 15+ years. MLOps is (model"  
[X Link](https://x.com/akshay_pachaar/status/2016854444445749311)  2026-01-29T12:42Z 246.9K followers, 32.7K engagements


"New research from NVIDIA. AI agents built an entire deep learning framework from scratch The project is called VibeTensor. It's an open-source PyTorch-style library with Python and Nodejs frontends backed by a C++20/CUDA core. Here's the crazy part: No human reviewed the code diffs. The agents proposed changes ran builds and tests and kept what passed. The entire system was built in approximately two months spanning 200K+ lines of code across tensors autograd CUDA memory management and custom kernels. One of the most interesting findings from this paper: The authors describe something called"  
[X Link](https://x.com/akshay_pachaar/status/2016930730367209790)  2026-01-29T17:45Z 247.4K followers, [----] engagements


"Link to the GitHub repo: (don't forget to star 🌟) https://github.com/microsoft/agent-lightning https://github.com/microsoft/agent-lightning"  
[X Link](https://x.com/akshay_pachaar/status/2017216720130441649)  2026-01-30T12:41Z 246.5K followers, [----] engagements


"Clawdbot/OpenClaw tutorial dropping tomorrow. What would make it useful for you"  
[X Link](https://x.com/akshay_pachaar/status/2017580079480717506)  2026-01-31T12:45Z 247.2K followers, [----] engagements


"@Everlier The idea is to understand how to build a powerful harness around the agent. I love Clawdbot and don't want to propose this as a substitute for it. nanobot codebase is easier to play with. thats it"  
[X Link](https://x.com/akshay_pachaar/status/2020182555530858879)  2026-02-07T17:07Z 247.5K followers, [----] engagements


"Anthropic released a 33-page guide on building Skills. Here's everything you need to know (under [---] words): First what are Skills A skill is a folder that teaches Claude how to handle specific tasks. You teach it once and it works every time. No more re-explaining your preferences in every conversation. Skills aren't locked to Claude. They've been published as an open standard so you can use them with AI agents like OpenClaw too. Here's the simplest way to think about it: MCP gives Claude access to your tools. Skills teach Claude how to use them well. One without the other is incomplete. The"  
[X Link](https://x.com/akshay_pachaar/status/2022220152910352763)  2026-02-13T08:03Z 248.2K followers, 142K engagements


"Download the PDF here: https://resources.anthropic.com/hubfs/The-Complete-Guide-to-Building-Skill-for-Claude.pdf https://resources.anthropic.com/hubfs/The-Complete-Guide-to-Building-Skill-for-Claude.pdf"  
[X Link](https://x.com/akshay_pachaar/status/2022220165354860680)  2026-02-13T08:03Z 248.2K followers, 28.6K engagements


"95% of AI engineering is just Context engineering. Everyone's obsessed with better models while context remains the real bottleneck. Even the best model in the world will give you garbage if you hand it the wrong information. Here's what most people miss: Context engineering isn't just about RAG or memory or agents. It's the art and science of delivering the right information in the right format at the right time to your LLM. Think about what you actually need: Retrieval to fetch relevant documents Short-term memory to track conversations Long-term memory to remember user preference Agents to"  
[X Link](https://x.com/akshay_pachaar/status/1991122721145823672)  2025-11-19T12:33Z 248K followers, 105.5K engagements


"Finally an open-source Python library for Context engineering Pixeltable is a unified declarative framework that handles your entire multimodal pipeline from data storage to model execution. The idea is simple: instead of stitching together a vector database a SQL database an embedding service and an agent framework everything lives in one system. Your documents embeddings conversation history and agent outputs are all just tables. Embeddings are computed columns that update automatically. Vector search works alongside your regular data operations. Built for end-to-end context engineering"  
[X Link](https://x.com/akshay_pachaar/status/1991409231040565379)  2025-11-20T07:32Z 247.9K followers, 67.2K engagements


"8 key skills to become a full-stack AI Engineer: (free/open-source resources below)"  
[X Link](https://x.com/akshay_pachaar/status/1994759678245712089)  2025-11-29T13:25Z 247.9K followers, 47.5K engagements


"In case you missed it. 💫 Yesterday I published a comprehensive video tutorial on fine-tuning Agents using RL. Reinforcement learning for LLMs always had one major problem: manually defining reward functions. You have to figure out how to score model outputs handle edge cases and tune everything until training works. This is where most people give up. ART (Agent Reinforcement Trainer) is an open-source framework that solves this with a simple approach: - Let the agent attempt tasks multiple times - An LLM judge relatively grades each attempt - The model learns from what worked vs what didn't"  
[X Link](https://x.com/akshay_pachaar/status/2016813331496259744)  2026-01-29T09:59Z 247.7K followers, 10.8K engagements


"Building AI agents dont read blogs dont watch videos write code and build things from scratch I've open-sourced 90+ projects on AI Agent RAG MCP and context engineering (27k+ stars already)"  
[X Link](https://x.com/akshay_pachaar/status/2017938839457952126)  2026-02-01T12:31Z 247.8K followers, 23.5K engagements


"this is a game-changer. Ollama just dropped: ollama launch a single command that sets up coding tools like Claude Code Opencode and Codex with local or cloud models. all you need is one command no environment variables or config files. think about what this means: - pick your model (local or cloud) - pick your tool - start coding that's it. want to run Claude code with a local glm-4.7-flash ollama launch claude done. the barrier to running AI-powered coding workflows just dropped to zero. https://twitter.com/i/web/status/2015365045623144561 https://twitter.com/i/web/status/2015365045623144561"  
[X Link](https://x.com/akshay_pachaar/status/2015365045623144561)  2026-01-25T10:04Z 248.2K followers, 56.7K engagements


"Microsoft did it again Building with AI agents almost never works on the first try. You spend days tweaking prompts adding examples hoping it gets better. Nothing systematic just guesswork. This is exactly what Microsoft's Agent Lightning solves. It's an open-source framework that trains ANY AI agent with reinforcement learning. Works with LangChain AutoGen CrewAI OpenAI SDK or plain Python. Here's how it works: Your agent runs normally with whatever framework you're using. Just add a lightweight agl.emit() helper or let the tracer auto-collect everything. Agent Lightning captures every"  
[X Link](https://x.com/akshay_pachaar/status/2017216707228848390)  2026-01-30T12:41Z 248.2K followers, 141.4K engagements


"Google just dropped another banger the figures in this paper were drawn by the system described in the paper. PaperBanana is an agentic framework that generates publication-ready academic illustrations from methodology descriptions. no manual design no Figma just your method section and a caption. here's how it works: five specialized agents collaborate in sequence: Retriever: finds relevant reference diagrams from a curated set of NeurIPS papers. matches by visual structure not topic. Planner: translates your methodology text into a detailed visual description using in-context learning."  
[X Link](https://x.com/akshay_pachaar/status/2018678023529357391)  2026-02-03T13:28Z 248.2K followers, 229.1K engagements


"http://x.com/i/article/2022272369113649152 http://x.com/i/article/2022272369113649152"  
[X Link](https://x.com/akshay_pachaar/status/2022309334483677654)  2026-02-13T13:58Z 248.2K followers, 271.5K engagements


"MiniMax-M2.5 is now fully open source. Think Opus [---] but 95% cheaper. Download model weights: GitHub: https://github.com/MiniMax-AI/MiniMax-M2.5 https://huggingface.co/MiniMaxAI/MiniMax-M2.5 https://t.co/TdpoOI1mqJ https://github.com/MiniMax-AI/MiniMax-M2.5 https://huggingface.co/MiniMaxAI/MiniMax-M2.5 https://t.co/TdpoOI1mqJ"  
[X Link](https://x.com/akshay_pachaar/status/2022574708051583120)  2026-02-14T07:32Z 248.2K followers, 111K engagements


"Nothing beats open-source MiniMax just dropped M2.1 and devs are calling it "Claude at 10% the cost." - 72.5% SWE-Multilingual. Beats Sonnet [---] - 88.6% VIBE-bench. Beats Gemini [--] Pro I used it to build an AI studio that turns any website into a podcast. 100% open-source"  
[X Link](https://x.com/akshay_pachaar/status/2003808497339924706)  2025-12-24T12:42Z 248.2K followers, 140.1K engagements


"My lecture at MIT✨ From Physics to Linear Algebra & Machine learning I have learned a lot from MIT Yesterday I had the honour of delivering a guest lecture on The state of AI Engineering exploring: - Prompt Engineering - Retrieval Augmented Generation. - Fine-Tuning Large Language Models. - And how LLMs are bringing a paradigm shift Thank you @RoyShilkrot for the invitation 🙏 Grateful for the opportunity to engage with such brilliant minds If you are interested AI Engineering I write a free weekly newsletter @ML_Spring"  
[X Link](https://x.com/akshay_pachaar/status/1724322731376918560)  2023-11-14T07:05Z 248.2K followers, 916.3K engagements


"Google just dropped a new LLM You can run it locally on just [---] GB RAM. Let's fine-tune this on our own data (100% locally):"  
[X Link](https://x.com/anyuser/status/1956334618715800005)  2025-08-15T12:38Z 248.2K followers, 2M engagements


"Make Claude Code 10x more powerful. Claude-Mem is a free plugin to persist memory across Claude sessions. It captures tool usage so you always start where you left off. Endless Mode allows 95% token reduction & 20x more tool use before context exhaustion. 100% open-source"  
[X Link](https://x.com/akshay_pachaar/status/2000181577905828176)  2025-12-14T12:30Z 248.2K followers, 186.2K engagements


"I switched to AI Engineering [--] years ago It was the best career move I ever made. If you want to start today here's a roadmap: . [--]. Master Python While many are skipping fundamentals for vibe coding those with strong coding foundations will always stand out. Python is the language the AI community speaks and Harvard's CS50p is the best place to learn it. 🔗 Harvard's CS50p [--]. Learn Vibe coding Here's the truth: You're not competing against AI you're competing against developers who use AI. Even if you're an excellent coder there are equally skilled developers shipping 3x faster because"  
[X Link](https://x.com/akshay_pachaar/status/2015044780330492116)  2026-01-24T12:51Z 248.2K followers, 257.6K engagements


"Set up Clawdbot/OpenClaw in under a minute. You don't need to buy a Mac mini use a terminal or worry about any config files. Here's what you can do: [--]. Go to the OpenClaw Lightning environment [--]. Click on "Clone." [--]. Follow the simple guide I've shared below From there you can connect it to WhatsApp or Telegram and control your assistant from your mobile. This turns [--] hours of installation and configuration into just [--] minutes. But there's something more important: This approach keeps everything secure sandbox environment off your local machine. Clawdbot is fun and powerful but I would never"  
[X Link](https://x.com/akshay_pachaar/status/2018328513015878038)  2026-02-02T14:19Z 248.2K followers, 47.9K engagements


"Everyone is sleeping on MiniMax Agent. It's like Clawdbot + Claude Cowork + Agent Skills combined. But more polished. I tested it by feeding an AI paper and asking it to create a presentation. I expected something basic. The output was actually usable. Good structure relevant visuals clean export. Watch the demo below. But presentations are just the start. Here's what else you can do with it: - Deep research via search browser use and MCPs - Build Full-stack apps with auth DB and Stripe - Connect to GitHub Figma Slack out of the box - Custom MCPs for your specific workflows - Multimodal"  
[X Link](https://x.com/akshay_pachaar/status/2019025725010133476)  2026-02-04T12:30Z 248.2K followers, 22.2K engagements


"OpenClaw/Clawdbot masterclass In this video I give you the complete blueprint to get up and running with your own Clawdbot. And not just one but multiple AI employees working 24x7 for you. Enjoy Chapters: 00:00 - Intro 1:12 - See my [--] AI employees in action 5:49 - How the whole thing works under the hood 9:47 - Installation and setup 11:22 - Hooking Clawdbot to Telegram 15:35 - Understanding the project structure 20:03 - Adding custom skill (from a repo of 42k skills) 23:18 - Automating your bots with Cron jobs 25:07 - Going from [--] to [--] Clawdbots 28:53 - Cron jobs vs. Heartbeats: what's the"  
[X Link](https://x.com/akshay_pachaar/status/2019515908960711122)  2026-02-05T20:58Z 248.2K followers, 148.6K engagements


"There's a pattern that keeps repeating in software. First everyone focuses on the building problem. Frameworks emerge mature and become genuinely good. Then suddenly the constraint flips. We saw this with neural networks. PyTorch and TensorFlow were excellent for building models. But deploying them meant dealing with different formats runtimes and infrastructure headaches. ONNX emerged to bridge that gap. We're watching the same pattern unfold with Agents right now. Frameworks like LangGraph CrewAI and LlamaIndex are mature enough that building an agent is no longer the hardest part. The hard"  
[X Link](https://x.com/akshay_pachaar/status/2019750445842149712)  2026-02-06T12:30Z 248.2K followers, 18.4K engagements


"The best AI projects are the ones you can actually understand. nanobot is an open-source personal AI assistant built in just [----] lines of code. That's 99% smaller than the project it's inspired by (Clawdbot has 430k+ lines). Most AI agent frameworks are bloated and impossible to navigate. nanobot takes the opposite approach and delivers core agent functionality with a codebase you can read on a weekend. Key features: - Works with any LLM provider - Run local models with vLLM - Connects to Telegram Discord and WhatsApp - Ships with scheduled tasks persistent memory and a skills system"  
[X Link](https://x.com/akshay_pachaar/status/2020137802302959665)  2026-02-07T14:09Z 248.2K followers, 55K engagements


"TinyLoRA: LoRA scaled down to [--] parameter. researchers from Meta Cornell and CMU just dropped a banger. they turned an [--] billion parameter model into a math and reasoning powerhouse by tweaking just [--] of those parameters. that's [--] bytes. less storage than this sentence. the model hit 91% accuracy on GSM8K up from 76% before the tweak. the method is called TinyLoRA and it pushes low-rank adaptation to its absolute extreme. some quick background on LoRA first: when you finetune a large model you're updating billions of parameters. LoRA showed you can instead learn a small low-rank update on"  
[X Link](https://x.com/akshay_pachaar/status/2020540105329639631)  2026-02-08T16:47Z 248.2K followers, 41.4K engagements


"Here's a common misconception about RAG Most people think RAG works like this: index a document retrieve that same document. But indexing retrieval. What you index doesn't have to be what you feed the LLM. Once you understand this you can build RAG systems that actually work. Here are [--] indexing strategies that separate good RAG from great RAG: 1) Chunk Indexing This is the standard approach. Split documents into chunks embed them store in a vector database and retrieve the closest matches. Simple and effective but large or noisy chunks will hurt your precision. 2) Sub-chunk Indexing Break"  
[X Link](https://x.com/akshay_pachaar/status/2021222919221936547)  2026-02-10T14:01Z 248.2K followers, 16.9K engagements


"TinyLoRA: LoRA scaled down to [--] parameter: TinyLoRA: LoRA scaled down to [--] parameter. researchers from Meta Cornell and CMU just dropped a banger. they turned an [--] billion parameter model into a math and reasoning powerhouse by tweaking just [--] of those parameters. that's [--] bytes. less storage than this sentence. https://t.co/F3bZRsshUV TinyLoRA: LoRA scaled down to [--] parameter. researchers from Meta Cornell and CMU just dropped a banger. they turned an [--] billion parameter model into a math and reasoning powerhouse by tweaking just [--] of those parameters. that's [--] bytes. less storage than"  
[X Link](https://x.com/akshay_pachaar/status/2021897353184325966)  2026-02-12T10:41Z 248.2K followers, 14.2K engagements


"MiniMax M2.5 is here - Open-source - SOTA on SWE bench - [----] 🔥 Introducing M2.5 an open-source frontier model designed for real-world productivity. - SOTA performance at coding (SWE-Bench Verified 80.2%) search (BrowseComp 76.3%) agentic tool-calling (BFCL 76.8%) & office work. - Optimized for efficient execution 37% faster at complex https://t.co/UwiKzzQNG8 Introducing M2.5 an open-source frontier model designed for real-world productivity. - SOTA performance at coding (SWE-Bench Verified 80.2%) search (BrowseComp 76.3%) agentic tool-calling (BFCL 76.8%) & office work. - Optimized for"  
[X Link](https://x.com/akshay_pachaar/status/2022003608532009321)  2026-02-12T17:43Z 248.2K followers, 10.3K engagements


"@_avichawla Great recommendation For anyone building with Claude or Cursor there's also an MCP server for RAG-Anything so you can plug multimodal document processing directly into Claude Desktop and other MCP clients as well"  
[X Link](https://x.com/akshay_pachaar/status/2022239667358458325)  2026-02-13T09:21Z 248.2K followers, [----] engagements


"Big moment for Postgres Search has always been Postgres' weak spot and everyone just accepted it. If you needed a real relevance-ranked keyword search the default answer was to spin up Elasticsearch or add Algolia and deal with the data sync headaches forever. The problem isn't that Postgres can't do text search. It can. But the built-in ts_rank function uses a basic term frequency algorithm that doesn't come close to what modern search engines deliver. So teams end up: - Running a separate Elasticsearch cluster just for search - Building sync pipelines that inevitably drift out of"  
[X Link](https://x.com/akshay_pachaar/status/2014314783391298019)  2026-01-22T12:30Z 248.2K followers, 217.8K engagements


"The most likely reason AI apps fail in production: (and it's not the LLM) Most teams only look at what goes in and what comes out. When the response is bad they blame the model and move on. But an AI pipeline has multiple steps and any one of them can silently break without you ever knowing. This is where observability comes in. Let me give you a mental model for how you must think about it: Take a RAG system as an example. A user asks a question it flows through multiple components and eventually a response comes out. Each step takes time each step can fail and each step has its own cost. To"  
[X Link](https://x.com/akshay_pachaar/status/2020482849506472097)  2026-02-08T13:00Z 248.2K followers, [----] engagements


"There's a new learning paradigm for AI agents. It learns the way humans do. Think about how you learned to drive. Nobody memorizes every route turn by turn. You develop instincts like maintaining safe distance anticipating what other drivers will do and braking early in the rain. Those instincts become skills you carry to every road you ever drive on. AI agents today do the opposite. Most memory systems store raw trajectories which are full logs of every action the agent took during a task. These logs are long noisy and full of redundant steps. Stuffing them into context often makes things"  
[X Link](https://x.com/akshay_pachaar/status/2021600955624722633)  2026-02-11T15:03Z 248.2K followers, 18.5K engagements


"Meta just solved the biggest problem in RAG Most RAG systems waste your money. They retrieve [---] chunks when you only need [--]. They force the LLM to process thousands of irrelevant tokens. You pay for compute you don't need. Meta AI just solved this. They built REFRAG a new RAG approach that compresses and filters context before it hits the LLM. The results are insane: - 30.85x faster time-to-first-token - 16x larger context windows - 2-4x fewer tokens processed - Outperforms LLaMA on [--] RAG benchmarks Here's what makes REFRAG different: Traditional RAG dumps everything into the LLM. Every"  
[X Link](https://x.com/akshay_pachaar/status/2022650095188262998)  2026-02-14T12:32Z 248.2K followers, [----] engagements


"GPT implemented using NumPy in like [--] lines of code 🔥 Let me break it down for to you in this thread: 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1627652574315954180)  2023-02-20T12:53Z 248.2K followers, 1.1M engagements


"I have been coding in Python for [--] years now. Here's a roadmap to Master Python 🚀 A Thread 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1637309911385407489)  2023-03-19T04:27Z 248.2K followers, 546.9K engagements


"Don't pay ridiculous amount of money to learn Python Machine Learning & Data Science. The likes of Google Microsoft & Top universities like Harvard Stanford & MIT are offering these for FREE 🙌 A Thread 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1639856037951070208)  2023-03-26T05:05Z 248.2K followers, 584.5K engagements


"Python is powerful 🔥 You can download a YouTube video in like [--] lines of code 🚀 Check this out 👇"  
[X Link](https://x.com/akshay_pachaar/status/1640332084026966016)  2023-03-27T12:36Z 248.2K followers, 353.8K engagements


"I have been coding in Python for [--] years now. Here's a roadmap to Master Python 🚀 A Thread 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1647479943755096065)  2023-04-16T06:00Z 248.2K followers, 417.2K engagements


"Microsoft is offering FREE courses in following areas: - AI - IOT - Data Science - Machine Learning A project-based pedagogy that allows you to learn while building 🚀 Read More 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1652289137737891841)  2023-04-29T12:30Z 248.2K followers, 407.2K engagements


"Harvard university is offering FREE world class education in Data Science Courses cover: - Python - Data Visualization - Probability - Statistics - Machine Learning - Data Science: Capstone A project-based pedagogy that allows you to learn while building 🚀 Read more 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1659899291610099713)  2023-05-20T12:30Z 248.2K followers, 552K engagements


"MIT offers free courses in following areas: - Python - Machine Learning - Deep Learning - Probability & Statistics - Linear Algebra A comprehensive curriculum where you get to learn from the best Read more 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1663885564687556611)  2023-05-31T12:30Z 248.2K followers, 775.3K engagements


"Google has created a Generative AI learning path with [--] FREE courses Topics cover: - Intro to LLMs - Attention Mechanism - Image Generation/Captioning - Intro to Responsible AI From the fundamentals of LLMs to creating & deploying generative AI solutions 🚀 Read more 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1665335120595922950)  2023-06-04T12:30Z 248.2K followers, 681.5K engagements


"Businesses need data analysis more than ever Microsoft offers a highly practical & hands-on roadmap to become a skilled Data Analyst 🧑🎓 You will learn: - Story telling - Data modeling - Data visualization - Decision making - Data analysis in Power BI A Thread 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1667509438364942338)  2023-06-10T12:30Z 248.2K followers, 491.8K engagements


"I have been coding in Python for [--] years now. Here are [--] amazing Python libraries not known to many A Thread 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1668596609490501632)  2023-06-13T12:30Z 248.2K followers, 563.7K engagements


"Stanford offers FREE Data Science Education World-class courses on: - Machine Learning - Computer Vision - Statistical Learning - Graph Neural Networks - Reinforcement Learning - Natural language processing (NLP) Read more 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1670046174206918656)  2023-06-17T12:30Z 248.2K followers, 578.3K engagements


"I've been coding in Python for [--] years now. If I were to start over again in [----] here's a roadmap: 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1673307658810687489)  2023-06-26T12:30Z 248.2K followers, 2.9M engagements


"Microsoft & LinkedIn just released a professional certificate in Generative AI It's FREE & comes in [--] course: - Introduction to AI. - What is Generative AI. - The evolution of Search. 🔍 - Streamline you work w/t Bing Chat. - Ethics in the Age of Generative AI. Read more 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1674757194107985921)  2023-06-30T12:30Z 248.2K followers, 631.1K engagements


"Harvard university is offering FREE world class education in Data Science Courses cover: - Python - Data Visualization - Probability - Statistics - Machine Learning - Data Science: Capstone A project-based pedagogy that allows you to learn while building 🚀 Read more👇"  
[X Link](https://x.com/anyuser/status/1680555390960844800)  2023-07-16T12:30Z 248.2K followers, 1.4M engagements


"Microsoft is offering FREE courses in following areas: - AI - IOT - Data Science - Machine Learning A project-based pedagogy that allows you to learn while building 🚀 Read More .👇"  
[X Link](https://x.com/anyuser/status/1684181230244495362)  2023-07-26T12:37Z 248.2K followers, 921.2K engagements


"Stanford offers FREE Data Science Education World-class courses on: - Machine Learning - Computer Vision - Statistical Learning - Graph Neural Networks - Reinforcement Learning - Natural language processing (NLP) Read more .👇"  
[X Link](https://x.com/anyuser/status/1693238967452459014)  2023-08-20T12:30Z 248.2K followers, 574.5K engagements


"Harvard university is offering FREE world class education in Data Science Courses cover: - Python - Data Visualization - Probability - Statistics - Machine Learning - Data Science: Capstone A project-based pedagogy that allows you to learn while building 🚀 Read more👇"  
[X Link](https://x.com/anyuser/status/1703031118663135623)  2023-09-16T13:00Z 248.2K followers, 652.6K engagements


"Python is powerful 🔥 You can download a YouTube video in like [--] lines of Code 🚀 Check this out 👇"  
[X Link](https://x.com/anyuser/status/1705499767382614122)  2023-09-23T08:30Z 248.2K followers, 660.3K engagements


"Arguably the best Linear Algebra course out there. It's Free Taught by MIT's legendary Professor Gilbert Strang. Build a strong foundation in mathematics for machine learning Check this out👇 https://www.youtube.com/playlistlist=PL221E2BBF13BECF6C https://www.youtube.com/playlistlist=PL221E2BBF13BECF6C"  
[X Link](https://x.com/anyuser/status/1710995972833378811)  2023-10-08T12:30Z 248.2K followers, 323.3K engagements


"I have been coding in Python for [--] years now. If I were to start over today here's a roadmap: 🧵👇"  
[X Link](https://x.com/anyuser/status/1713110082093989916)  2023-10-14T08:30Z 248.2K followers, 758.7K engagements


"Linear Regression clearly Explained Linear Regression models relationship between a dependant variable (y) & two or more independent variables (x1 x2 .) ❗For the sake of simplicity we discuss linear regression with a single independent variable. Mathematical Representation: y=mx+c where: - y is the dependent variable - x is the independent variable - m is the slope of the line - c is the y intercept Cost Function: ❗The goal is to find the best values for m and c that minimize the error between the predicted values & the actual values for all data points. This is how the cost function looks a"  
[X Link](https://x.com/anyuser/status/1714619920955346995)  2023-10-18T12:30Z 248.2K followers, 337.7K engagements


"Microsoft is offering FREE courses in following areas: - AI - IOT - Data Science - Machine Learning A project-based pedagogy that allows you to learn while building 🚀 Read More .👇"  
[X Link](https://x.com/anyuser/status/1721433502212895093)  2023-11-06T07:45Z 248.2K followers, 478.2K engagements


"I have been coding in Python for [--] years now. Here are [--] amazing Python libraries not known to many.👇"  
[X Link](https://x.com/anyuser/status/1726216300652003384)  2023-11-19T12:30Z 248.2K followers, 1.2M engagements


"I started my career in Data Science back in [----] Here's a list of Tech YouTube Channels I've ardently followed.👇"  
[X Link](https://x.com/anyuser/status/1728752968181453108)  2023-11-26T12:30Z 248.2K followers, 603.5K engagements


"This Python library is so powerful 🔥 Introducing pix2tex 🚀 Now you can convert images of equations into LaTeX in just [--] lines of code👇 from PIL import Image from pix2tex.cli import LatexOCR img = Image. open('path/to/image.png') model = LatexOCR() print(model(img)) Repo: If you are interested in AI Engineering subscribe to my technical newsletter @ML_Spring (5k+ readers): https://mlspring.beehiiv.com/subscribe https://github.com/lukas-blecher/LaTeX-OCR https://mlspring.beehiiv.com/subscribe https://github.com/lukas-blecher/LaTeX-OCR"  
[X Link](https://x.com/anyuser/status/1733826773388009554)  2023-12-10T12:31Z 248.2K followers, 256.4K engagements


"I have been coding in Python for [--] years now. If I were to start over in [----] here's a roadmap: 🧵👇"  
[X Link](https://x.com/akshay_pachaar/status/1735638370033016925)  2023-12-15T12:30Z 248.2K followers, 776.1K engagements


"Eigenvalues & Eigenvectors clearly explained The concept of eigenvalues & eigenvectors is widely known yet poorly understood Today I'll clearly explain their meaning & significance. Let's go 🚀"  
[X Link](https://x.com/akshay_pachaar/status/1738175074174406922)  2023-12-22T12:30Z 248.2K followers, 408.3K engagements


"How LLMs work clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1755207285880803610)  2024-02-07T12:30Z 248.2K followers, 904.1K engagements


"I recently landed my job as an AI Engineer Here's a roadmap for those starting today:"  
[X Link](https://x.com/anyuser/status/1762817455184052407)  2024-02-28T12:30Z 248.2K followers, 1.1M engagements


"Eigenvalues & Eigenvectors clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1773326676342239308)  2024-03-28T12:30Z 248.2K followers, 1.2M engagements


"Microsoft has launched one of the most comprehensive course on Generative AI This course is [--] lessons where each lesson covers its own topic so start wherever you would like Lessons are labeled either "Learn" lessons explaining a Generative AI concept or "Build" lessons that explain a concept and code examples in both Python and TypeScript when possible. Here's the link to the courses: If you're interested in AI Engineering I write a weekly Newsletter @ML_Spring Join 8k+ reader now: https://mlspring.beehiiv.com/subscribe https://github.com/microsoft/generative-ai-for-beginners"  
[X Link](https://x.com/anyuser/status/1774051586844791203)  2024-03-30T12:30Z 248.2K followers, 407.1K engagements


"5 GitHub repositories that will give you superpowers as an AI/ML Engineer:"  
[X Link](https://x.com/akshay_pachaar/status/1777314795827737041)  2024-04-08T12:37Z 248.2K followers, 578.2K engagements


"Let's build a RAG app using MetaAI's Llama-3 (100% local):"  
[X Link](https://x.com/anyuser/status/1781299233162965149)  2024-04-19T12:30Z 248.2K followers, 1.1M engagements


"How LLMs work clearly explained:"  
[X Link](https://x.com/anyuser/status/1788184595621716204)  2024-05-08T12:30Z 248.2K followers, 815.6K engagements


"Object oriented programming in Python clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1802680778377347334)  2024-06-17T12:32Z 248.2K followers, 617.7K engagements


"Eigenvalues & Eigenvectors clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1806304022967804045)  2024-06-27T12:30Z 248.2K followers, 792.4K engagements


"Let's build a RAG app using MetaAI's Llama-3.1 (100% local):"  
[X Link](https://x.com/anyuser/status/1816088785152848028)  2024-07-24T12:31Z 248.2K followers, 713.3K engagements


"Neural network visualized"  
[X Link](https://x.com/anyuser/status/1819713171739041913)  2024-08-03T12:33Z 248.2K followers, 580.5K engagements


"I've been coding in Python for [--] years now. If I were to start over today here's a roadmap:"  
[X Link](https://x.com/anyuser/status/1838918900538552389)  2024-09-25T12:30Z 248.2K followers, 399.8K engagements


"Web scraping will never be the same Crawl4AI simplifies web crawling and data extraction making it ready to use for LLMs and AI applications. Heres why its a game-changer: Free and open-source 🚀 Blazing fast performance 🤖 LLM-friendly output formats (JSON cleaned HTML markdown) 🌍 Supports crawling multiple URLs simultaneously 🎨 Extracts all media tags (Images Audio Video) 🔗 Extracts all external and internal links But thats not all: 📚 Extracts metadata from pages 🕵 User-agent customization 🖼 Takes screenshots of pages 📜 Executes custom JavaScript before crawling Link to the GitHub"  
[X Link](https://x.com/anyuser/status/1841463917060735224)  2024-10-02T13:03Z 248.2K followers, 256.5K engagements


"Bayes theorem clearly explained:"  
[X Link](https://x.com/anyuser/status/1845102716420297086)  2024-10-12T14:02Z 248.2K followers, 761.5K engagements


"Microsoft just changed the game 🔥 They've open-sourced bitnet.cpp: a blazing-fast 1-bit LLM inference framework that runs directly on CPUs. Why is this a game-changer❓ You can now run 100B parameter models on local devices with up to 6x speed improvements and 82% less energy consumptionall without a GPU The future we've been waiting for: fast efficient and private AI that works anytime anywhere.✨ Link to the GitHub repo in next tweet _____ Find me @akshay_pachaar✔ For more insights and tutorials on AI and Machine Learning"  
[X Link](https://x.com/anyuser/status/1847312752941085148)  2024-10-18T16:24Z 248.2K followers, 805.4K engagements


"Stanford CS229: Building Large Language Models This [---] hours lecture provides a concise overview of building a ChatGPT-like model covering both pretraining (language modeling) and post-training (SFT/RLHF). For each component it explores common practices in data collection algorithms and evaluation methods. Link in the next tweet _____ Find me @akshay_pachaar ✔ For more insights and tutorials on AI and Machine Learning"  
[X Link](https://x.com/anyuser/status/1852690165321658829)  2024-11-02T12:32Z 248.2K followers, 235.5K engagements


"SAMURAI vs. MetaAI's SAM [--] Traditional visual object tracking struggles in crowded fast-moving or self-occluded scenes as does SAM2. Meet SAMURAI: a completely open-source adaptation of the Segment Anything Model for zero-shot visual tracking Here's why it's a game-changer: 🚫 No need for retraining or finetuning 🎯 Boosts success rate and precision 🤖 Motion-aware memory selection 💪 Zero-shot performance on diverse datasets But that's not all: 🔬 Refines mask selection 🔮 Predicts object motion effectively 📈 Gains: 7.1% AUC on LaSOT 3.5% AO on GOT-10k 🏆 Competes with fully supervised"  
[X Link](https://x.com/anyuser/status/1859937514691371031)  2024-11-22T12:30Z 248.2K followers, 369.4K engagements


"Turn any GitHub repository into LLM-ready text Simply replace "hub" with "ingest" in a GitHub URL and receive a prompt-friendly text ingest for LLMs. Gitingest is 100% open-source and provides: - Directory structure - A brief summary of the project - The entire content as LLM-ready text Plus it comes with a nice python package and you can run the UI locally Stay tuned I'm working on something really cool with this✨ Link to the GitHub repo in the next tweet _____ Find me @akshay_pachaar ✔ For more insights and tutorials on ML and AI Engineering"  
[X Link](https://x.com/anyuser/status/1871171889810620568)  2024-12-23T12:31Z 248.2K followers, 197.5K engagements


"FireDucks makes Pandas 125x Faster. .by changing JUST ONE LINE of code. Pandas has some major limitations: - Pandas only uses a single CPU core. - It often creates memory-heavy DataFrames. - Its eager (immediate) execution prevents global optimization of operation sequences. FireDucks is a highly optimized drop-in replacement for Pandas with the same API. To use FireDucks: Remove this: import pandas as pd Add this: import fireducks.pandas as pd And you're done ✔ A performance comparison of FireDucks vs. DuckDB Polars and Pandas is shown in the video below. Official benchmarks indicate: -"  
[X Link](https://x.com/anyuser/status/1871534809383891024)  2024-12-24T12:34Z 248.2K followers, 224.4K engagements


"I've been coding in Python for [--] years now. If I were to start over today here's a roadmap:"  
[X Link](https://x.com/anyuser/status/1874432889267569055)  2025-01-01T12:30Z 248.2K followers, 411K engagements


"Microsoft has released its own document parser for LLM use . . Introducing MarkItDown a 100% open-source one-stop solution for effortlessly converting any file to Markdownperfect for text analysis indexing and more Heres what makes it special: Converts PDF Word Excel PPT images audio to markdown Extracts EXIF OCR and transcripts automatically Available via CLI Python API or Docker Offers LLM-based image descriptions Supports batch conversions Link to the repo in next tweet _____ Find me @akshay_pachaar ✔ For more insights & tutorials on AI and Machine Learning"  
[X Link](https://x.com/anyuser/status/1874795304920445123)  2025-01-02T12:30Z 248.2K followers, 412.6K engagements


"Web scraping will never be the same Crawl4AI simplifies web crawling and data extraction making it ready to use for LLMs and AI applications. Heres why its a game-changer: Completely free and open-source 🚀 Blazing fast performance outperforming many paid services 🤖 LLM-friendly output formats (JSON cleaned HTML markdown) 🌍 Supports crawling multiple URLs simultaneously 🎨 Extracts all media tags (Images Audio Video) 🔗 Extracts all external and internal links But thats not all: 📚 Extracts metadata from pages 🔄 Custom hooks for auth headers and page modifications 🕵 User-agent"  
[X Link](https://x.com/anyuser/status/1875166693154271300)  2025-01-03T13:05Z 248.2K followers, 214.9K engagements


"Let's build a RAG app using DeepSeek-R1 (100% local):"  
[X Link](https://x.com/anyuser/status/1881741968507838861)  2025-01-21T16:33Z 248.2K followers, 718.3K engagements


"Let's build a multi-agent YouTube video analyst powered by DeepSeek-R1 (100% local):"  
[X Link](https://x.com/anyuser/status/1882406721974624278)  2025-01-23T12:35Z 248.2K followers, 513.3K engagements


"Let's compare DeepSeek-R1 and OpenAI-o1 using RAG:"  
[X Link](https://x.com/anyuser/status/1883497754649047192)  2025-01-26T12:50Z 248.2K followers, 1.1M engagements


"How LLMs work clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1887117394109423890)  2025-02-05T12:33Z 248.2K followers, 1.1M engagements


"Microsoft has launched a powerful new data analysis tool Introducing Data Formulator a 100% open-source LLM-powered no-code tool that transforms data in a snap and creates stunning visualizations. Key features include: 🤖 AI-powered data transformation 🖱 Interactive drag-and-drop UI for visualizations 💬 Seamless blend of UI & natural language inputs But thats not all: You can even create charts beyond your initial dataset. Data Formulator automatically identifies extra computation needs generates fields for you and outputs the final visualization. Find the GitHub repo in the next tweet"  
[X Link](https://x.com/anyuser/status/1891832889228263729)  2025-02-18T12:51Z 248.2K followers, 288.3K engagements


"Model Context Protocol (MCP) clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1900170356494917936)  2025-03-13T13:01Z 248.2K followers, 1.1M engagements


"Model Context Protocol (MCP) clearly explained: Model Context Protocol (MCP) clearly explained: Model Context Protocol (MCP) clearly explained:"  
[X Link](https://x.com/anyuser/status/1900456268084764689)  2025-03-14T07:57Z 248.2K followers, 342.9K engagements


"I just built my own OCR app powered by Google's Gemma [--]. It runs 100% locally on my computer using Ollama and extracts text from images as structured markdown. The code is open-source and you can find it in the next tweet"  
[X Link](https://x.com/anyuser/status/1901249805571252580)  2025-03-16T12:30Z 248.2K followers, 487.5K engagements


"Microsoft open-sourced a no-code data analysis tool Introducing Data Formulator an LLM-powered no-code tool that transforms data in a snap and creates stunning visualizations using AI. 100% open source"  
[X Link](https://x.com/anyuser/status/1902701376993738834)  2025-03-20T12:38Z 248.2K followers, 314.7K engagements


"GitHub's Official MCP Server is here You can use it to automate GitHub workflows analyze repository data and build AI tools that interact with GitHub's ecosystem. 100% Open-source"  
[X Link](https://x.com/akshay_pachaar/status/1908497370683506959)  2025-04-05T12:30Z 248.2K followers, 201.2K engagements


"MCP is on fire. AI agents can now talk to real world tools apps and actually get stuff done. This changes everything. Here are [--] amazing examples:"  
[X Link](https://x.com/anyuser/status/1910671710300627006)  2025-04-11T12:30Z 248.2K followers, 721.7K engagements


"MCP vs A2A (Agent2Agent) protocol clearly explained:"  
[X Link](https://x.com/anyuser/status/1912123161883345321)  2025-04-15T12:37Z 248.2K followers, 975.4K engagements


"10 MCP AI Agents and RAG projects for AI Engineers (with code):"  
[X Link](https://x.com/akshay_pachaar/status/1912856065135591891)  2025-04-17T13:09Z 248.2K followers, 487.4K engagements


"INCREDIBLE Gemini did [--] hours of manual data analysis work in just [--] mins. Gemini in a Colab notebook now lets you plan analyze and visualize your data. I used a dataset of 100k rows and didn't write or execute a single line of code. Completely hands-off"  
[X Link](https://x.com/akshay_pachaar/status/1913575832284758298)  2025-04-19T12:50Z 248.2K followers, 265.5K engagements


"How LLMs work clearly explained:"  
[X Link](https://x.com/akshay_pachaar/status/1919368985092780189)  2025-05-05T12:30Z 248.2K followers, 762.2K engagements


"Traditional RAG vs. Agentic RAG clearly explained (with visuals):"  
[X Link](https://x.com/anyuser/status/1920818537968660949)  2025-05-09T12:30Z 248.2K followers, 840.9K engagements


"Microsoft just changed the game 🔥 They've open-sourced bitnet.cpp: a blazing-fast 1-bit LLM inference framework that runs directly on CPUs. Why is this a game-changer❓ You can now run 100B parameter models on local devices with up to 6x speed improvements and 82% less energy consumptionall without a GPU The future we've been waiting for: fast efficient and private AI that works anytime anywhere.✨ The model is also available on HuggingFace you can try it right away I've shared Link to the GitHub repo in next tweet where you'll find all the details. _____ Find me @akshay_pachaar ✔ For more"  
[X Link](https://x.com/anyuser/status/1921909676733583854)  2025-05-12T12:45Z 248.2K followers, 527.6K engagements


"Let's build an MCP-powered financial analyst (100% local):"  
[X Link](https://x.com/anyuser/status/1927372063259377984)  2025-05-27T14:31Z 248.2K followers, 567.1K engagements


"Model Context Protocol (MCP) clearly explained:"  
[X Link](https://x.com/anyuser/status/1933503158225154514)  2025-06-13T12:34Z 248.2K followers, 1.1M engagements


"I switched to AI Engineering [--] years ago It was the best career move I ever made. If you want to start today here's a roadmap:"  
[X Link](https://x.com/anyuser/status/1954158220727263311)  2025-08-09T12:30Z 248.2K followers, 2.2M engagements


"Traditional RAG vs. Agentic RAG clearly explained (with visuals):"  
[X Link](https://x.com/anyuser/status/1955245388241440925)  2025-08-12T12:30Z 248.2K followers, 688.1K engagements


"Everyone is sleeping on this new OCR model dots-ocr is a new 1.7B vision-language model that achieves SOTA performance on multilingual document parsing. - Supports 100+ languages - Works with both images and PDFs - Handles text tables formulas seamlessly 100% open-source"  
[X Link](https://x.com/anyuser/status/1956695061162561815)  2025-08-16T12:30Z 248.2K followers, 259.7K engagements


"JSON prompting for LLMs clearly explained:"  
[X Link](https://x.com/anyuser/status/1957784212112830798)  2025-08-19T12:38Z 248.2K followers, 1.1M engagements


"A simple technique makes RAG up to 40x faster & 32x memory efficient - Perplexity uses it in its search index - Google uses it in Vertex RAG engine - Azure uses it in its search pipeline Let's understand how to use it in a RAG system (with code):"  
[X Link](https://x.com/anyuser/status/1958510665217532012)  2025-08-21T12:45Z 248.2K followers, 523K engagements


"After MCP A2A & AG-UI there's another Agent protocol. It's fully open-source and launched by IBM Research. Here's a complete breakdown (with code):"  
[X Link](https://x.com/anyuser/status/1959601448402251899)  2025-08-24T12:59Z 248.2K followers, 468.2K engagements


"A team at DeepMind wrote this piece on how you must think about GPUs. Essential for AI engineers and researchers:"  
[X Link](https://x.com/anyuser/status/1966177460984348709)  2025-09-11T16:30Z 248.2K followers, 175.9K engagements


"Bayes theorem clearly explained:"  
[X Link](https://x.com/anyuser/status/1966863679007441302)  2025-09-13T13:56Z 248.2K followers, 167.9K engagements


"Microsoft did it again Building with AI agents almost never works on the first try. You spend days tweaking prompts adding examples hoping it gets better. Nothing systematic just guesswork. This is exactly what Microsoft's Agent Lightning solves. It's an open-source framework that trains ANY AI agent with reinforcement learning. Works with LangChain AutoGen CrewAI OpenAI SDK or plain Python. Here's how it works: Your agent runs normally with whatever framework you're using. Just add a lightweight agl.emit() helper or let the tracer auto-collect everything. Agent Lightning captures every"  
[X Link](https://x.com/anyuser/status/1984236460510519631)  2025-10-31T12:30Z 248.2K followers, 206.3K engagements


"A Data Science/ Machine Learning Roadmap 🚩🚀 🧵 👇"  
[X Link](https://x.com/akshay_pachaar/status/1566768581244702720)  2022-09-05T12:41Z 248.2K followers, [----] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing