At Paperlyft, I built a production AI platform that converts scanned PDFs into accessible documents while preserving their content, layout, and geometry. I developed multimodal workflows for page generation, validation, and repair, with accessibility checks and human feedback built into the process.
At ASU Learning Engineering Institute, I built Critical Code Reviewer, a GitHub Action that uses an LLM to review pull requests and post feedback on relevant code lines. I improved its reviews with structured prompts, repository context, and evaluation of useful versus noisy findings.
At Braincuber Technologies, I progressed from junior engineer to owning development of an AI healthcare initiative, working with doctors and patient consultation transcripts. I also helped turn a hackathon prototype into a live platform for NMTSA, and built Vctrx, an agentic AI coding harness with role-based context, tools, memory, and human approval.
