AI/ML Engineer building production AI systems with LLMs, agents, RAG, and backend infrastructure.
I enjoy turning AI capabilities into reliable software — from multi-agent workflows and retrieval systems to APIs, asynchronous workers, observability, and deployment.
- LLM & Agentic Systems — Multi-agent workflows, tool-using agents, structured generation, and orchestration
- RAG & Knowledge Systems — Hybrid retrieval, embeddings, vector databases, and context optimization
- AI Backend Systems — FastAPI services, asynchronous workers, APIs, databases, and distributed workloads
- Production AI Infrastructure — Docker, observability, evaluation/telemetry pipelines, and scalable deployments
Autonomous regression testing platform combining AST-based dependency analysis, 3-layer hybrid retrieval, 11 specialized LangGraph agents, sandboxed test execution, and automated failure diagnosis.
TestPilot AI analyzes code changes, identifies regression impact, retrieves relevant repository context, generates tests, executes them in isolated environments, diagnoses failures, and automates the review workflow.
Stack: Python · FastAPI · LangGraph · Tree-sitter · Qdrant · PostgreSQL · Celery · Redis · Docker · Next.js
- Production-grade LLM and agentic architectures
- RAG optimization and repository-scale retrieval
- AI evaluation and observability
- Distributed AI workloads and asynchronous systems
- Building reliable AI products rather than isolated model demos
Building AI systems that are useful, reliable, and production-ready.