Writing · IntelligenceOS
I publish the engineering work.
Notes on what I'm building, evaluating and getting wrong across production AI, financial systems, model adaptation and architecture.
Subscribe on Substack ↗
Latest · 90 seconds
The Model Is Not the AI System
The model is probably not the hardest part of your AI system. The hard part is everything around it: data, tools, state, permissions, evaluation, failure recovery and verification.
Watch on YouTube ↗AI evaluationProduction agentsFinancial automationPayment systemsArchitectureRAGFine-tuningReliabilityFailure analysis
INTELLIGENCEOSThe Reframe: You’re Not Learning AIThe build is the curriculum. Choose a real friction, make one testable loop, and let failure write the syllabus.2026-08-21Read ↗INTELLIGENCEOSThe IOS Agent Layer — Building Reliable Multi-Step Analysis Pipelines for FinanceA production agent layer for document analysis is not a GPT wrapper with function calling. It’s a state machine with failure handling, a typed tool registry, deterministic routing…2026-06-19Read ↗INTELLIGENCEOSDomain Embedding Fine-Tuning — Adapting BGE to Finance and Risk DocumentsThis post addresses that gap. Not by switching models, but by teaching the one we have.2026-06-18Read ↗INTELLIGENCEOSGeneral Embeddings Are a Tax on Domain AIToday’s deep dive covers exactly this: how we built BGE-finance-v1, what the training data looks like, what Matryoshka fine-tuning gives you, and what a 24-point nDCG improvement…2026-06-08Read ↗INTELLIGENCEOSHow to Design APIs for AI Agentic SystemsAn API that an agent can call is not the same as an API an agent can use correctly.2026-06-05Read ↗INTELLIGENCEOSAgents Don’t Need to Be Smart — They Need to Be ReliableThe autonomy narrative conflates two things: capability and reliability. A highly capable agent can do many things. A reliable agent does a smaller set of things consistently and…2026-05-25Read ↗INTELLIGENCEOSThe IOS Architecture — A Complete Blueprint for a Domain AI PlatformThis post is the synthesis. How all of these components fit together into a coherent production system for a specific domain. Not a generic architecture diagram — a real system…2026-05-11Read ↗INTELLIGENCEOSEnterprise AI’s Real Moat Is Domain Data, Not Model QualityIn a market where model quality is converging toward commodity, access to good models is not a moat. It’s table stakes...what does this mean for you and your company?2026-05-05Read ↗