Weights & Biases: the AI developer platform. Build better models faster, fine-tune LLMs, develop GenAI applications with confidence, all in one system of record developers are excited to use. W&B Models is the MLOps solution used by foundation model builders and enterprises who are training, fine-tuning, and deploying models into production. W&B Weave is the LLMOps solution for software developers who want a lightweight but powerful toolset to help them track and evaluate LLM applications. Weights & Biases is trusted by over a 1,000 companies to productionize AI at scale including teams at OpenAI, Meta, NVIDIA, Cohere, Toyota, Square, Salesforce, and Microsoft. Sign up for a 30-day free trial today at http://wandb.me/trial.
External link for Weights & Biases
Weights & Biases helps AI developers build better models faster. Quickly track experiments, version and iterate on datasets, evaluate model performance, reproduce models, and manage your ML workflows end-to-end.
400 Alabama St
San Francisco, California 94110, US
Weights & Biases reposted this
Finally, Main Event Day! Countless attendees are showing us their passionate will! AI! Easy to demo, But Hard to production! Check it out with Weights & Biases !
Weights & Biases reposted this
We have some big news to share: Marimo is joining CoreWeave! We’re doubling down on open-source and scaling molab with serious compute. Our mission remains the same: to build the world’s best open-source programming environment for working with data. To our community: we wouldn’t be here without your enthusiasm, feedback, and contributions. We’re so grateful — and so excited to keep building with you. marimo is, and always will be, free, open-source, and permissively licensed. Read the full announcement: https://lnkd.in/g6cfFRJa
We are less than a week away from Fully Connected London! Fully Connected lands Nov 4–5 at Convene Sancroft, St. Paul’s. Two days focused on one outcome: move agentic AI from prototype to production. What you’ll do: • Bring a laptop. Leave with a working agentic pipeline, instrumented with evals, tracing, and alerting. • Build with mentors from the W&B team and take home a sample repo, checklists, and dashboards. • Learn the playbooks teams use in production: eval suites that catch regressions, dataset hygiene, safety gates, failover, caching, and cost-aware routing. What to expect: • Day 1: hands-on workshops. • Day 2: speakers and live demos with candid case studies. • Short talks, long hallway conversations, and space to whiteboard with peers. Seats are moving fast. Details and tickets here → https://lnkd.in/g9tJcygN
Weights & Biases reposted this
🚀 Day 1 of Weights & Biases Fully Connected Tokyo 2025 has finally begun from Yesterday's Welcom Reception Dinner. While Day 2 will be held tomorrow, the event kicks off today with a customer workshop featuring several AI companies. We've seen a tremendous turnout from day one, with a huge number of customers showing keen interest. We're looking forward to Day 2 tomorrow. LFG!!! APAC Team! Upstage Si Yoo Keisuke Kamata Satoshi Isemura Masaya Shigematsu Richard Y. Song Yuya Yamamoto Hyunwoo Oh Lukas Biewald Chris Van Pelt Seann Gardiner Robin Bordoli
Weights & Biases reposted this
Sales Development Team Lead at Weights and Biases I Deep Learning | MLOps | Generative AI/LLMs + Prompt Engineering
If you missed us yesterday at NVIDIA GTC DC make sure to make your way over to booth #435 and visit the Weights & Biases bee army 🪄🐝 #NVIDIA #GTC #MLops #MachineLearning #DC
Traditional AI benchmarks and their complex scores can be abstract and hard to interpret. How do you really know which model is performing better in a way anyone can understand? The team at ShiftLayer is building an exciting answer: BrainPlay. It’s a decentralized AI benchmark that ditches static tests and instead makes models compete in games. Their first implementation uses Codenames, allowing people to visually understand an AI's performance by watching it play. To make this competitive benchmark traceable and transparent, they just integrated Weights & Biases. As AI models compete, the system's validators now log all game outcomes, model performance metrics, and reward distributions straight to W&B. This provides a clear, live system of record for every game played, making their entire open ecosystem verifiable. It's a fantastic approach to making AI evaluation more engaging and understandable. See how they're building it and contribute to their open project on GitHub: https://lnkd.in/dgupT5S2
When you're building AI for healthcare, "good enough" can be dangerous. Hallucinations aren't just bugs; they are serious safety risks. So how do you prove your system is grounded, citable, and ready for production? This is why Simon (Xi) Ouyang's new guide on building a production ready healthcare RAG is such a must read. Evaluation wasn't just an afterthought for him; it was the core of his process. He used the Weights & Biases Evaluation Suite for continuous quality assurance, which is brilliant. Here’s how he used W&B: - Tracked custom metrics that actually matter in healthcare: think grounding_rate and citation_rate, not just generic scores. - Logged complete audit trails using wandb. Tables, capturing every single question and answer. This is huge for compliance. - Used W&B as a quality gate in his automated workflow, preventing bad models from ever reaching production. - Compared experiments side by side to show exactly how he improved his model's grounding rate by 20%! This is an amazing blueprint for building safe, auditable, and reliable AI. Read the full technical deep dive by Simon Ouyang here: wandb.me/xiouyang
Weights & Biases reposted this
Wow! With a week to go, we have 900+ people registered for Weights & Biases Fully Connected in London on 4-5 November! 🤯 Still time to lock in your attendance! Here's what to expect... ️⚡️Hands‑on workshops for building safe, robust AI agents (Day 1) 🎤 Real-world stories from AI builders and leaders at organisations such as Sanofi, Shell, GSK, Mistral AI, NVIDIA, Wayve, LlamaIndex, AI Security Institute & many more (Day 2) 📊 Explore best practices for operationalizing machine learning and GenAI systems 🤝 Network with like-minded AI practitioners, exchanging experiences and hard-won lessons in building AI for real-world application We very much look forward to welcoming you 🙌 Registration link in comments...
We've launched a complete overhaul of our Weights & Biases docs! All wandb product docs are now under one roof to create a unified experience that is faster, smarter, and more connected for our community. What’s new? - Unified Platform: One central hub for all W&B products. - Powerful AI Search: An AI chat assistant and "Ask AI" search are integrated across every doc to get you answers instantly. (Built with W&B Weave) - Modern Experience: Real-time search, live translations, and a clean, React-powered design. - Developer-Focused: We've added OpenAPI support and even in-IDE search capabilities. This update provides seamless, intuitive navigation and integrated AI helpers that boost productivity, from finding information in the docs to writing your code. This move allows us to focus 100% on what matters most, which is improving content and shipping new features for our users!