I build systems
where humans
stay in control.
Production LLM systems designed around real product constraints, not just model capabilities. Augmenting people, not replacing them.
Content Ingestion for AI-Native Learning — SQL Walkthrough
Bounded prototype walkthrough showing how partial SQL teaching material can be transformed into a structured, inspectable learning graph with critique, repair, clarification, and deterministic validation.
↗LLM agents and production systems at GriffinAI
End-to-end architecture and delivery of production agentic LLM products — Transaction Execution Agent, Cardano Proposal Examiner, and multi-agent ops.
↗Knowledge Assistant — Trust-Critical Retrieval Across Fragmented Internal Knowledge
Permission-aware internal knowledge assistant focused on trust-critical retrieval, citations, access control, and calibrated uncertainty across fragmented company knowledge.
Technology should augment people, not replace them.
I'm not a tech evangelist or a luddite. The question isn't how fast we automate — it's how we evolve with technology without losing ourselves.
Start from constraints
User needs first, model is a detail.
Make the workflow explicit
State, routing, evaluation, observability.
Keep humans in control
Automation where it earns its place.
Exocore
A personal context system — capturing inputs, structuring them into tasks, memory, and decisions. Not autonomous. Human-in-the-loop by design.