Turn AI potential
into working systems
An independent AI lab exploring how modern AI systems actually get built and deployed. We share open research, run hands-on education programs, and geek out on the stuff that matters in production.
What We Do
Built from real production experience
Three areas we dig into β research, education, and the engineering that makes AI actually work in the wild.
AI Education Programs
Hands-on workshops and open curricula that advance AI literacy β helping engineering, product, and leadership teams build confidently with AI.
Get involved βRAG Systems Research
Digging deep into retrieval-augmented generation β embedding strategy, chunking, retrieval evaluation β and sharing everything we find openly.
Get involved βAgentic AI Research
Exploring how AI agents actually behave in production β what makes them reliable, where they break, and how to scope them for real workflows without the drama.
Get involved βLatest Thinking
From the blog
Claude Mythos 5 + Record VC Boom + Tesla AI5 Tape-Out β April 20, 2026
Anthropic's Claude Mythos 5 marks the first 10-trillion-parameter model reaching production, arriving as Q1 2026 shatters all venture funding records with $300B
April 19, 2026: Mythos + GPT-5 + Record VC Reshape the AI Stack
April 2026 marked the single densest month for frontier AI model releases in history. Anthropic shipped Claude Mythos (exclusive to 50 partner organizations thr
Opus 4.7 β April 18, 2026
AnthropiΡ released Claude Opus 4.7, its most powerful publicly available model, dominating every hardest coding benchmark while simultaneously launching Claude
Let's Talk
Building something with AI?
Whether you want to swap notes on a production problem, explore an education program, or just geek out on what's working in AI β I'm always up for a good conversation.