I'm an AI Developer / Software Engineer with 4+ years building production, cloud-native applications — Node.js/Express & Java/Spring Boot microservices on Kubernetes, Angular/React front-ends, and Python data pipelines. Today I own ELI end-to-end — a conversational-AI assistant deployed across Florida's medical-board sites — serving 100,000+ users a month at 99.9%+ uptime.
Alongside the day job I go deep on applied AI: RAG, multi-agent systems, LLM fine-tuning (LoRA/QLoRA/ORPO), and the evals and guardrails that make them trustworthy. I don't just prototype models — I ship and operate them.
I pivoted from civil engineering into software, formalized it with an M.S. in Computer Science, and have been building ever since.
Short, high-intensity builds — where several of the applied-AI projects above came from.
Applied-AI systems I build outside the day job — ownership labeled honestly. Click through for the full case study.
NL→SQL clinical agent: AST SQL guardrails, QLoRA+ORPO fine-tuning, a 1,200-line eval harness.
case study →Graph-native longitudinal memory for clinical agents — versioned fact nodes, path-traversal retrieval, refuses to guess.
case study →Supply-chain risk monitor — I owned 7 live data feeds, LLM cost controls, and the 10-tab dashboard.
case study →Agentic CFO assistant (ADK + MCP + Gemini) with a numeric-provenance guard against hallucinated figures.
case study →PyTorch attention U-Net (CBAM+ASPP) for imbalanced disaster-imagery segmentation.
case study →Public source for the projects here — plus more as I open repos up.
github.com/pbiyyani09 →