At Powell Industries, I’m spearheading an AI defeaturing pipeline to automate CAD-to-simulation transfer from Autodesk Inventor to COMSOL. I also authored the financial analysis and product management plan that secured funding for the initiative.
As an R&D Software Engineering Intern at Powell Industries, I productionized a multi-agent RAG pipeline in Snowflake for all five R&D product lines. It processed more than 200 queries per day with 98% accuracy, and I delivered agent-to-agent communication, automated monitoring, and a self-improvement pipeline eight weeks early.
In Texas A&M’s SUCCESS Lab, I’m researching multi-robot coordination with an LLM agent architecture using NVIDIA Nemotron and OpenClaw. I also architected a security layer addressing prompt injection, context manipulation, and cross-layer exploitation in robotics environments.
At Texas A&M’s DIGIT Lab, I’m co-authoring research on interpretable machine learning for mechanical engineering and directed development of a symbolic regression pipeline to identify feature interactions. In other projects, I built an affective AI framework for humanoid robots and an open-source MCP server for Canvas LMS.
