Backend engineer with 4 years in fintech — I build systems that move money reliably at scale. My work spans distributed event pipelines, async workflow orchestration, and AI integration in production environments.

I've built semantic search engines from scratch (pgVector + OpenAI embeddings), priority-driven event systems processing 15k events/min inspired by Meta's FOQS, and piloted SLM-based personalization (Qwen, Gemma) in production engagement flows. On the infrastructure side: multi-stage workflow orchestration with Camunda Zeebe, atomic state transitions with Redis + Lua scripting, and async event pipelines that convert blocking flows into sub-15s non-blocking ones.

I care about correctness under concurrency, clean system boundaries, and shipping things that don't break at 2am.

• Currently building Graph Harness — a persistent semantic code graph for autonomous agents and CI systems
• Go-curious, Kafka-native, Redis-literate

Core

Data

AI / ML

Cloud & DevOps

Graph Harness active · Go

Persistent semantic layer for codebases. Combines event sourcing, graph traversal, and static analysis to build a deterministic, versioned Unified Code Graph — designed for humans, CI systems, and autonomous coding agents. Kernel + pluggable layer architecture with selector-driven resolution and replayable event logs.

_{Building things that don't break at 2am.}