At OpenEvidence, I architected a multi-agent framework for medical evidence synthesis and engineered a hybrid RAG pipeline that achieved 96% retrieval precision. I also optimized LLM inference throughput by 4x and developed a real-time clinical voice AI pipeline.
At Intellias, I designed LLM-powered assistants and contextual retrieval systems that automated up to 60% of manual operational workflows. I also developed MLOps pipelines that accelerated release cycles by 40%.
At UiPath, I developed AI components for RPA, including document understanding, OCR, and computer vision. I reduced document processing time by 50% and improved extraction accuracy by 35% through workflow optimization.
At KNAPP, I worked on full-stack applications and machine learning systems for search, voice assistance, fraud detection, forecasting, and review classification. My work included a fraud detection system that reduced false positives by 15% and forecasting models that improved in-stock rates by 12%.

