AI Engineer — RAG, agent workflows, and LLM evaluation
I build practical LLM systems with a software engineering foundation: retrieval-grounded applications, agent workflows, evaluation tooling, and reliability improvements for AI agent infrastructure. 5 years of software development experience across backend, full-stack, and applied AI engineering.
Retrieval-grounded resume Q&A app with a multi-model evaluation pass. Deployed on Hugging Face Spaces; evolved into an open-source LLM evaluation toolkit with parallel judge scoring and structured scoring outputs.
Reusable evaluation pipeline for LLM outputs: prompt-versioned scoring rubrics, parallel judge calls, and reliability checks. Built from the evaluation needs of the resume agent project.
Real-time interview assistant: local speech-to-text, streaming model responses, system-audio capture, and screen context tagging. Runs fully on local hardware.
Building RAG systems, LLM evaluation tooling, and AI-agent infrastructure projects with public artifacts.
Java/Spring services, multithreaded batch data loaders, MySQL schema work, and Linux/GCP operations for real-estate data integrations.