Senior Engineer · AI Engineering · Backend/Platform · Agent-Driven Development

Building LLM-powered systems that remember, reason, and ship. Author of virage (RAG pipeline), kms (agent knowledge management), and edap_model (civilization dynamics modeling).

TypeScript/Rust · 1.2M+ LOC indexed · Production-grade

The RAG pipeline that indexes once and stays in sync with every commit. Powers agent memory for AI coding assistants. Built with incremental AST-aware indexing, hybrid search (semantic + keyword), and eval-driven quality gates.

Stack: TypeScript, Rust, Tree-sitter, PostgreSQL/pgvector, GitHub Actions

Shell/Python · Plugin for Claude Code, Codex, Kilo

Gives AI coding agents durable, checkable knowledge — decisions, facts, and guardrails they won't forget, contradict, or re-litigate. Implements ADR/decision records, fact extraction, guardrail enforcement, and skill composition.

Stack: Shell, Python, Markdown, Git

Python/TeX · GPU-accelerated · Calibrated on Rome, USSR, Russia, USA

Mathematical model of civilization dynamics simulating how elite resource extraction erodes cooperation and technology, trapping societies in feudal equilibrium. Addresses Fermi Paradox via Great Plateau/Filter hypothesis.

Stack: Python (JAX/CUDA), TeX, NumPy/SciPy, Matplotlib

TypeScript Python Rust LangChain LlamaIndex Tree-sitter pgvector Eval-driven Dev

C#/.NET 8 Node.js Go PostgreSQL DynamoDB Redis Kafka/Kinesis gRPC REST

AWS (ECS, Lambda, RDS, DynamoDB) Azure Terraform GitHub Actions Kubernetes Docker

Angular 21 React TypeScript Tailwind PrimeNG Next.js

Prometheus Grafana ELK OpenTelemetry xUnit Jest Playwright

Agent-Driven Development — AI agents as first-class team members, not autocomplete. Durable context > prompt engineering.

Knowledge as Infrastructure — Decisions, facts, and guardrails are versioned, reviewable, and enforced — same as code.

Pragmatic Craft — Ship iteratively. Measure what matters. Optimize for change. Technical excellence serves product outcomes.

Systems Thinking — Local optimizations fail. Model the whole: data flow, team topology, incentive structures, failure modes.

• Taxonomy Is a Language — Why AI agents need structured vocabulary, not just more context (Medium, Sep 2026)
• The Project I Inherited Had No Memory Either — Onboarding failures as knowledge management gaps (DEV, Sep 2026)
• Two Kinds of Memory — Episodic vs semantic memory in agent systems (DEV, Sep 2026)
• My AI Keeps Forgetting What We Already Decided — The case for durable decision records (DEV, Sep 2026)
• Building a Terraform Provider for Paddle Billing — Real-sandbox testing strategies (DEV, Sep 2026)

All DEV articles → · All Medium articles →

• Designed & delivered 20+ microservices on .NET 8, Python, Node.js across AWS (ECS, Lambda, DynamoDB, RDS) & Azure
• Integrated Generative AI into internal tooling: RAG pipelines, content generation, AI-coding-agent memory systems
• Built CI/CD & IaC from scratch (GitHub Actions + Terraform); technical leadership via architecture POCs, code reviews, mentoring

• Built high-load RESTful APIs & microservices on AWS (.NET Core, EF Core, Dapper)
• Optimized PostgreSQL/RDS: 2–3× query performance improvement
• Designed async messaging flows (ActiveMQ, SQS, Kinesis)

• Led 4–9 engineers building infrastructure for cryptocurrency mining operations
• Architected custom OS distribution for GPU farms + high-performance ETH pool software
• Stack: Azure, Kubernetes, .NET Core, Kafka, Redis, MongoDB, ClickHouse, PostgreSQL, ELK, Prometheus, Grafana, Angular

Open to: Senior/Staff IC roles · AI Engineering leadership · Technical advisory · Agent-driven development consulting