I build the full stack of intelligent systems: the agent harnesses, memory layers, self-correcting loops, and production infrastructure that make AI reliable at scale.
My focus: agentic architectures that plan, execute, and course-correct autonomously. Subagent orchestration. Context engineering for token efficiency. RLHF pipelines that close the feedback loop.
Research in LLM safety, cloud security, and low-resource NLP. Published in IEEE conferences.
Agent Harness & Autonomous Loops
Custom harnesses with tool orchestration, subagent spawning, and self-correcting loops. No human in the loop.
Context Engineering & Memory
Context-compact architectures, RAG pipelines, and persistent memory for coherent multi-turn agents.
RLHF & Eval Pipelines
Human-in-the-loop annotation, automated evaluation, iterative prompt improvement. Data-driven agent quality.
Published Researcher
IEEE conference publications. Ongoing work in LLM safety and low-resource NLP.
