Homen Shum

AI Engineer & Technical Founder with deep experience in LLMs, multi-agent systems, and real-world AI product delivery. I specialize in building production-grade generative AI, RAG pipelines, and autonomous agents—driving impact across healthcare, finance, and e-commerce. Career Highlights: Generative AI & LLMs: Building with LLM APIs and generative tools since 2022. Launched CelebrityVoiceAI.com (GPT-3, Flask) to replicate celebrity speech patterns using vector search and real-time chatbot memory. Cybersecurity & Semantic Search: Won 2nd place at the Nuharbor Cybersecurity Hackathon (2023) for developing a GPT-powered semantic alert pairing engine, achieving 86%+ cross-vendor accuracy. Healthcare AI: Led FluencyMed to top-25 at the UC Berkeley AI Hackathon—real-time transcription and medical code mapping (CPT/ICD) using RAG. Became technical co-founder at CosmaNeura, launching a HIPAA-compliant clinical automation platform on Azure. AI in Finance: Architected LLM-powered automation for JPMorgan’s healthcare banking division—reducing data processing for 2,000+ companies from two weeks to 30 seconds using GPT, LlamaIndex, and Pydantic. Multi-Agent & RAG Systems: Currently building ParselyFi and CaféCorner—multi-agent platforms integrating web search, document ingestion, and structured report generation for business, medical, and event discovery use cases. Recent AI Web Apps: Medical Code Search & Report Generation Patient Screening Chat & Report Generation Auto Shop E-Commerce RAG Chat App Prospecting Assistant App Let’s connect if you’re passionate about AI agents, LLM workflows, or scaling real-world automation!