At Acclivousbyte, I build production AI features for a healthcare-staffing platform and an enterprise IBM-partner chatbot. For Assist ATS, I architected an LLM query router that cut unnecessary LLM calls by 15% and latency by 9%, and built a Salesforce agent with a self-correcting retry loop.
For FalconStor Thomas, I built a six-intent workflow router and an end-to-end PDF RAG pipeline with Elasticsearch and IBM WatsonX. I also implemented centralized activity logging with an admin dashboard and CSV export.
Outside those platforms, I independently built and deployed four full-stack AI SaaS products, including CrossBorder AI Law, a jurisdiction-aware legal RAG assistant, and Aperture, a personalized AI news platform. My work spans retrieval, agent workflows, and AI products built with Python, FastAPI, and Next.js.
