AI Engineer — prompt systems, agentic AI
I build reliable LLM workflows: evaluation pipelines, and agent automation
Background in ML data quality and LLM evaluation
- system-prompt-benchmark — security testing across attack vectors (injection, jailbreaks, leakage)
- dspy-optimization-patterns — teacher-student optimization patterns for quality/cost trade-offs
- RAG Agent — with hybrid search, page-number citations, and a LangGraph state machine
- Production RAG Pipeline — free Perplexity-style search for local sLLM
- llmflow-search — deep research agent that synthesizes reports from multiple web sources
Python TypeScript LangGraph DSPy RAG Prompt Engineering Agentic AI