Excited to share my latest Gen AI project – StyleAI: AI-Powered Fashion Design Generator!While building StyleAI, my goal was to explore how Generative AI can make fashion recommendations more personalized, creative, and accessible.Here's how the system works:👗 Google Gemini AI powers intelligent outfit recommendations based on user preferences such as style, occasion, budget, and season.💻 Python & Streamlit provide a clean and interactive interface, enabling users to receive fashion suggestions in real time.⚡ Prompt Engineering is used to generate tailored recommendations, ensuring that every output aligns with the user's specific requirements.The project demonstrates how AI can transform traditional fashion discovery into a more personalized and engaging experience while showcasing the practical applications of Generative AI in solving real-world problems.A special thank you to Dr.Nanthini Mohan Mohan for her guidance and mentorship throughout this journey. Your sessions provided a strong foundation in AI concepts and inspired me to apply my learning through hands-on projects like this one.This project was developed as part of my learning journey during the Edunet Foundation AI Internship Program.

Code Llama 4 from ScratchThis lecture teaches how to code Llama 4 from scratch.Contents - Introduction to the course- Llama 4 Overview and Ranking- Course Prerequisites- Course Approach for Beginners- Why Code Llama from Scratch?- Understanding LLMs and Text Generation- How LLMs Predict the Next Word- Probability Distribution of Next Words- The Role of Data in Prediction- Probability Distribution and Word Prediction- Sampling Techniques- Greedy Sampling- Random Sampling- Top K Sampling- Temperature Sampling for Controlling Randomness- What are Tokens?- Tokenization Example: "Hello world"- How LLMs Learn Semantic Meaning- Token Relationships and Context- The Concept of Embeddings- Tokenization Challenges- Large Vocabulary Size- Handling Misspellings and New Words- Introducing Subword Tokens- Byte Pair Encoding (BPE) Overview- Understanding Vector Embeddings- Visualizing Embeddings- The Embedding Layer- Token Indexing and Swapping Embeddings- Coding Your Own Tokenizer- Implementing Byte Pair Encoding- Initializing Vocabulary and Pre-tokenization- Splitting Text into Words- Calculating Pair Frequencies- Merging Frequent Pairs- Updating Vocabulary and Tokenization Rules- Implementing the Merges- Encoding Text with the Tokenizer- Decoding Tokens Back to Text- Self-Attention Mechanism- Query, Key, and Value Vectors- Calculating Attention Scores- Applying Softmax- Weighted Sum of Values- Self-Attention Matrix Operations- Multi-Head Attention- Implementing Self-Attention- Masked Self-Attention- Rotary Positional Embeddings (RoPE)- Understanding Positional Information- How RoPE Works- Implementing RoPE- Feed-Forward Networks (FFN)- Linear Layers and Activations- Implementing FFN

Hello connections✨ Generative AI refers to a category of artificial intelligence algorithms and models that generate new content, such as images, text, audio, or video, based on patterns and examples in the data they are trained on. These models use techniques like deep learning and neural networks to understand and replicate the data distribution they were trained on, allowing them to create new, original content that resembles the input data.✨ AIMER Society - Artificial Intelligence Medical and Engineering Researchers Society#AIMERS#AI#text#audio#techInnovation

Everyone is sleeping on this new OCR model!(open-source, 91 languages, < 1B params)Datalab just released Surya OCR, which is just a state-of-the-art OCR model scoring 83.3% on the olmocr benchmark (top under 3B)- 650M params- supports 91 languages- 5 pages/s on RTX 5090- runs on CPU, GPU, MPS- 83.3% olmocr bench score (top under 3B)- Full layout information- Extracts + captions images and diagrams- Strong handwriting, math, form, table support

7 open source repos that do what cursor, perplexity, and v0 do combined zero waitlists. zero api markups. zero vendor lock in. save the list: » openhands // agentic coding platform. plans, edits files, runs terminal commands, browses the web on its own. mit https://lnkd.in/gmxhbTCH » zed // rust-native ai editor. gpu-accelerated, ships the open agent client protocol so any agent can plug in. 86k stars https://lnkd.in/g_w75viP » codex // openai's official terminal coding agent. runs in network-isolated containers for secure work. 94k stars, apache 2.0 github.com/openai/codex » scira (formerly miniperplx) // minimalist perplexity alternative. citations, focus modes, built on vercel ai sdk. 11.7k stars https://lnkd.in/g7KdmvQg » morphic // open source answer engine with generative ui. streams charts and comparisons inline with the answer. mit github.com/miurla/morphic » december // self-hosted bolt.new. $5 in api credits gets you 50+ full app generations. docker deploy in one command https://lnkd.in/gihiCf7B » libra // cloudflare-native v0 and lovable alternative. serverless, free tier friendly, byok https://lnkd.in/gVg9cPyU closed source is losing this race quietly. star them before your feed catches on.

384GB of VRAM. On your desk. Not in a datacenter. Autonomous, the company behind SmartDesk, just open sourced complete build guides for a personal AI computer that runs open models on hardware no one can switch off. It's called Autonomous Computer. Three configurations, one philosophy: a model you rent can be cut off overnight. A model running in your own house can't. → Home build: 2x RTX 5090 workstation → Team build: 4x RTX PRO 6000 Blackwell, 384GB VRAM → On-prem business build: 8x RTX 4090/5090, up to 256GB VRAM Every config ships with the full bill of materials, 3D printable and CNC housing files, wiring diagrams, BIOS tuning, and assembly photos. Software side covers OS setup, NVIDIA drivers, and serving open models through Ollama, vLLM, and llama.cpp. Fork it. Change it. Build it. Sell it. MIT License. 100% Opensource. Repo: https://lnkd.in/gA6z_E4k

You are very lucky if this is showing up on your feed before 4th July. 𝐃𝐨 𝐲𝐨𝐮 𝐡𝐚𝐯𝐞 𝐚 𝐡𝐚𝐫𝐝 𝐭𝐢𝐦𝐞 𝐠𝐞𝐭𝐭𝐢𝐧𝐠 𝐚 𝐫𝐞𝐟𝐞𝐫𝐫𝐚𝐥? 𝐈’𝐯𝐞 𝐜𝐫𝐚𝐜𝐤𝐞𝐝 𝐭𝐡𝐞 𝐫𝐞𝐚𝐥 𝐡𝐢𝐫𝐢𝐧𝐠 𝐡𝐚𝐜𝐤. 𝐀𝐟𝐭𝐞𝐫 𝐜𝐨𝐮𝐧𝐭𝐥𝐞𝐬𝐬 𝐚𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬, 𝐞𝐧𝐝𝐥𝐞𝐬𝐬 𝐢𝐧𝐭𝐞𝐫𝐯𝐢𝐞𝐰 𝐜𝐚𝐥𝐥𝐬, 𝐚𝐧𝐝 0 𝐣𝐨𝐛 𝐨𝐟𝐟𝐞𝐫𝐬. ❌ It’s not your resume ❌ It’s not your LinkedIn headline ❌ It’s not your “open to work” badge ✅ It’s your connections/networking I’ve prepared a sheet that includes all the openings along with LinkedIn profiles from Startups to MAANG and emails of more than 4500+ recruiters. If you are really looking for job Checkout: https://lnkd.in/gJ6gRzDN Or 💬 𝐂𝐨𝐦𝐦𝐞𝐧𝐭 "𝐒𝐇𝐄𝐄𝐓" 𝐚𝐧𝐝 𝐈'𝐥𝐥 𝐬𝐞𝐧𝐝 𝐲𝐨𝐮 𝐭𝐡𝐞 𝐬𝐡𝐞𝐞𝐭 𝐝𝐢𝐫𝐞𝐜𝐭𝐥𝐲 𝐭𝐨 𝐲𝐨𝐮𝐫 𝐃𝐌! #sde1 #sde2 #sde3 #sde4 #hiring #jobHunt #explore

I'm sooo bullish on open source AI. Was paying $15/month for a popular AI voice to text tool. But switched to an open source one where you download the model to your computer, and everything is done locally. It's better, faster, more private, and it's free! I wonder what other subscriptions I can get rid of?

Everyone's building a company brain. Almost nobody understands what that actually means.Karpathy's personal LLM knowledge base went viral. Garry Tan built GBrain. YC people are clearly paying attention to the category.But here's what most of these get wrong:A personal brain is small, fast, and permission-less. A company brain serves hundreds of people, enforces access controls, and has to heal itself when it goes wrong.Most tools are just personal brains scaled up. They break at the seams.Slite's thesis is different: → Company knowledge needs a verification layer → It needs to be legible to both humans and agents simultaneously → It needs to self-correct, not just storeAfter 10 years building knowledge bases and 2 years on AI search, they just shipped the first 𝐬𝐞𝐥𝐟-𝐦𝐚𝐢𝐧𝐭𝐚𝐢𝐧𝐢𝐧𝐠 𝐤𝐧𝐨𝐰𝐥𝐞𝐝𝐠𝐞 𝐛𝐚𝐬𝐞.The Slite Agent monitors 20+ connected tools, detects drift, proposes fixes, and queues them for human review. Claude, ChatGPT, Cursor - any MCP-compatible agent, can read and write to it directly.Accurate context at the company level is a different problem than accurate context at the personal level.Someone finally built for the right problem.The interesting part isn't that Slite added AI to docs.It's that they're treating company knowledge like infrastructure

Turn any GitHub repo into an RL training environment!Repo2RLEnv is an open-source tool from HuggingFace that synthesizes verifiable RL training data from any GitHub repository. Point it at any repo and it automatically generates tasks, verifies them, and pushes datasets directly to HuggingFace Hub.The core idea: every merged PR in any GitHub repo is a solved problem. Someone found a bug, fixed it, and the tests passed. Repo2RLEnv turns thousands of these real solved problems into structured RL training tasks automatically.The pipeline is end-to-end. It bootstraps a Docker environment for the repo, mines merged PRs, creates tasks with the broken codebase as the starting state, verifies each task passes the quality gate (tests fail before the fix, tests pass after), and pushes the final dataset to HuggingFace Hub.The reward is verifiable. Agent attempts a fix, tests run in the sandbox, pass means +1, fail means 0. No LLM judge needed for most pipelines.Nine synthesis pipelines cover different scenarios. PR diffs for fast text-only tasks. PR runtime for sandbox-verified real bugs. Mutation bugs for artificially injected failures. CVE patches for real security vulnerabilities. Equivalence tests for function-level synthesis. Each pipeline flows through the same QA gate before tasks are admitted.Works with any coding agent harness for evaluation. Claude Code, OpenHands, Codex CLI, Gemini CLI all supported out of the box via Harbor compatibility.Key capabilities:• Converts any GitHub repo into verifiable RL training tasks• 9 synthesis pipelines covering PRs, commits, bugs, CVEs, and refactors• Automated Docker bootstrap per repo, cached after first run• Verifiable rewards via test execution or diff similarity• Pushes datasets directly to HuggingFace Hub• Compatible with TRL, SkyRL, Prime-RL and other RL trainers• Works with private repos end-to-end

Your iPhone can now run the largest local AI models. Borrowing compute from your MacBook 🤯LM Studio just shipped LM Link for iPhone.100% free. Works with every Open Source model you can run locally.The problem with local AI has always been hardware.Your MacBook or home server can run massive open models. Your phone can't. So the moment you leave your desk, you're back to cloud APIs and subscriptions.LM Link fixes that.Your models keep running on your Mac at home. Your iPhone connects to them securely from anywhere. You chat with your biggest local models from a coffee shop, a flight gate, the back of a cab.What you get:• Full access to your largest models, right from your phone• End-to-end encryption between your devices• All chats saved locally on your devices, nothing touches a cloud server• Works through the Locally app, now the official LM Studio mobile appSetup takes minutes. Install LM Studio on your Mac, turn on LM Link, download Locally on your iPhone, pair the devices.That's it. Your phone becomes a window into your home AI rig. Private AI used to mean being chained to your desk. Now your hardware travels with you.

Your iPhone can now run the largest local AI models. Borrowing compute from your MacBook 🤯LM Studio just shipped LM Link for iPhone.100% free. Works with every Open Source model you can run locally.The problem with local AI has always been hardware.Your MacBook or home server can run massive open models. Your phone can't. So the moment you leave your desk, you're back to cloud APIs and subscriptions.LM Link fixes that.Your models keep running on your Mac at home. Your iPhone connects to them securely from anywhere. You chat with your biggest local models from a coffee shop, a flight gate, the back of a cab.What you get:• Full access to your largest models, right from your phone• End-to-end encryption between your devices• All chats saved locally on your devices, nothing touches a cloud server• Works through the Locally app, now the official LM Studio mobile appSetup takes minutes. Install LM Studio on your Mac, turn on LM Link, download Locally on your iPhone, pair the devices.That's it. Your phone becomes a window into your home AI rig. Private AI used to mean being chained to your desk. Now your hardware travels with you.

This shouldn't be free. Someone open sourced a fully local AI meeting notetaker that kills Otter, Fireflies, and Fathom in one Rust app. It's called Meetily. Every meeting gets recorded, transcribed, and summarized entirely on your own machine. Zero cloud. Zero API calls to a company you don't trust with your audio. The pitch isn't just privacy. It's speed too. → Whisper and Parakeet models both bundled in, 4x faster live transcription → Ollama handles summaries locally, or plug in Claude, Groq, OpenRouter, or any OpenAI-compatible endpoint → Captures your mic and system audio at the same time with intelligent ducking so nothing clips → GPU acceleration ships automatic. Metal and CoreML on Apple Silicon. CUDA on NVIDIA. Vulkan on AMD and Intel. → Import any old recording and re-transcribe it with a better model or a different language → Native app for macOS and Windows. Linux builds from source. The average data breach costs $4.4M. Every cloud notetaker on the market is a compliance risk waiting to fire. Meetily is what happens when engineers finally build the alternative. 13.2K stars. MIT License. 100% Opensource. https://lnkd.in/geyf7rsR

 the tool, its open-source MIT release, the speculative decoding mechanism, and the general performance direction are all confirmed.

 this is essentially a verbatim reproduction of a real, published letter from a credible, named source (Andrew Ng via DeepLearning.AI's The Batch), with accurate context on the "loop engineering" term's origin.

Someone built a tool that turns any codebase into an interactive knowledge graph you can explore, search, and ask questions about.It's called Understand Anything.Here's what actually happens when you run it =>Six agents analyze your project in parallel. A knowledge graph gets built of every file, function, class, and dependency. An interactive dashboard opens, color-coded by architectural layer, fully searchable, every node clickable with plain-English explanations.Then you get =>→ /understand-chat - ask anything about the codebase in plain English → /understand-diff - see what your changes affect before committing → /understand-explain - deep dive into any specific file or function → /understand-onboard - generate an onboarding guide for new team members → /understand-domain - extract business domain knowledge as a horizontal graph → /understand-knowledge - analyze a wiki and surface implicit relationships as a navigable idea graphThe persona-adaptive UI is a detail I hadn't seen before, the dashboard adjusts its detail level depending on whether you're a junior dev, PM, or power user.It supports multilingual output too. Run /understand --language zh and the entire dashboard generates in Chinese.It works with Claude Code, Cursor, Copilot, Gemini CLI, Codex, OpenCode, Vibe CLI, Cline, KIMI CLI.

every technical claim — built-in PII types, the three custom-detector methods, and the compliance use cases — is directly sourced from LangChain's own documentation and code.

Don't train the model, evolve the harness.I read a brilliant blog post from Hugging Face where they took a frozen open model scoring 0% on a hard legal agent benchmark, left its weights alone, and let an automated loop rewrite only the code around it.That code layer is the harness, the runtime wrapper that feeds the model context, runs its tool calls, and decides when a run ends.By the time the loop finished, the system had essentially matched Sonnet 4.6 on the benchmark's headline metric, at roughly 7x lower cost per task. Zero weights changed.The gain existed because of where the model was failing. The judge only grades files saved in the right place under the exact requested filename, and the model kept doing the legal analysis correctly, then saving it under the wrong name, dropping it in a scratch folder, or never writing it at all.So the 0% was never measuring legal reasoning. It was measuring the harness.Hand-tuning that layer is slow and model-specific, so they automated it. A Claude proposer adds exactly one mechanism per iteration, and an outer loop keeps it only if it clearly beats the current best, so accepted mechanisms compound.What the loop discovered says a lot about where agents actually fail.→ The biggest single gain was file handling, not intelligence. An automatic step that lands the deliverable exactly where the judge expects it beat every prompt change, with zero extra model tokens.→ Code fixes transferred across models, prompt playbooks did not. The same harness lifted a smaller model from the same family by 14 points, but the tuned prompts hurt a different model family on tasks it could already finish.→ The harness mattered more than anything else. Same model, same judge, same tasks, and five different harnesses scored anywhere between 3.5% and 80.1%.The gains do eventually flatten, and the remaining misses look like real capability gaps. At some point the wrapper runs out of tricks and the model has to carry the work.But the lesson holds. A benchmark score measures the model and its harness together, and until the harness is fixed, it's impossible to know which one failed.I highly recommend reading the article. Link in the first comment.

Quick recap: every specific feature (worktree isolation, Design Mode, SSH worktrees, GitHub/Linear integration, mobile companion, scriptable CLI, MIT license) is independently verified across multiple sources. 

NVIDIA 𝐢𝐧𝐭𝐫𝐨𝐝𝐮𝐜𝐞𝐝 𝐍𝐯𝐢𝐝𝐢𝐚 𝐓𝐰𝐨 𝐓𝐨𝐰𝐞𝐫, 2𝐱 𝐟𝐚𝐬𝐭𝐞𝐫 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥Diffusion language models offer a promising alternative to autoregressive models due to their potential for parallel and iterative generation. 𝐈𝐬𝐬𝐮𝐞 𝐰𝐢𝐭𝐡 𝐞𝐱𝐢𝐬𝐭𝐢𝐧𝐠 𝐝𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥𝐬 However, existing approaches use a single network for both context representation and iterative denoising, forcing one model to serve both roles and limiting its capacity for either role. 𝐓𝐰𝐨 𝐓𝐨𝐰𝐞𝐫 𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥TwoTower, a block-wise autoregressive diffusion model that decouples these roles into two towers: ✅ a frozen AR context tower that causally processes clean tokens, and ✅ a trainable diffusion denoiser tower with bidirectional block attention that refines noisy blocks via cross-attention to the context. Built on Nemotron-3-Nano-30B-A3B, an open-weight 30B hybrid Mamba-Transformer MoE model, and trained on approximately 2.1T tokens. Nemotron-Labs-TwoTower retains 98.7% of the autoregressive baseline's quality while offering 2.42X higher wall-clock generation throughput.