At KeeMakr.ai, I architected and deployed BTU Virtual University, a production multi-agent platform serving 500+ concurrent student sessions. Its Agentic RAG pipeline delivered contextual responses in under two seconds across 11 agents, with 99.5% production uptime.
As an independent contractor, I built Squatch Signal, a lead intelligence pipeline that turns company websites into scored, evidence-backed prospects. I also designed its Supabase backend and deployed the Streamlit front end.
Earlier, I delivered AI and Generative AI solutions spanning RAG, document intelligence, and AI assistants, and developed production AI models at Huly.ai. I transitioned from clinical physiotherapy into AI engineering through structured technical coursework, certifications, and hands-on production deployments.

