I build production-grade AI systems — not demos. Every project ships with authenticated APIs, observable infrastructure, Azure Bicep IaC, CI/CD pipelines, and documentation designed for client handoff.
| Project | What it solves | Stack |
|---|---|---|
| LLM Observatory | Real-time monitoring for hallucination risk, latency, token cost, and eval drift across LLM-powered features | FastAPI · Prometheus · App Insights · Azure OpenAI |
| Enterprise RAG | Citation-backed document Q&A with role-based access control enforced at the retrieval layer | FastAPI · Azure AI Search · Entra ID · PostgreSQL |
| Document Intelligence Workbench | Structured extraction from invoices and contracts — confidence scoring, rule validation, human review queue | FastAPI · Azure Document Intelligence · Blob Storage |
| ProcureAI | Five specialist agents validate purchase requests against policy, budget, vendor risk, and approval authority | FastAPI · LangChain · MCP · A2A · PostgreSQL |
| Multi-Agent Orchestration | Enterprise workflow automation with persistent human-in-the-loop approvals and tamper-evident audit trails | LangGraph · FastAPI · SQLAlchemy · Anthropic · OpenAI |
→ View full portfolio with screenshots and briefs
LLM Observability ████████████████████ monitoring · eval · alerting · cost tracking
Enterprise RAG ████████████████████ retrieval · access control · citations · eval
Document Intelligence ████████████████████ extraction · validation · review · audit trail
Multi-Agent Systems ████████████████████ LangGraph · MCP · A2A · human-in-the-loop
Azure Infrastructure ████████████████████ Container Apps · Key Vault · AI Search · Bicep
Every project I deliver meets this baseline — not as a checklist, but as a starting point:
- Auth — JWT, API keys, OIDC, Azure Entra ID
- Observability — Prometheus metrics, Application Insights, structured JSON logs, Langfuse LLMOps tracing
- Infrastructure — Azure Bicep IaC, Container Apps, Key Vault, Managed Identity, no hardcoded secrets
- CI/CD — GitHub Actions: lint → test → build → deploy-staging → deploy-prod with manual gate
- Data — PostgreSQL with Alembic migrations, role-based access, audit logging
- Testing — pytest, integration tests, golden eval datasets per project
Available for fixed-scope delivery, custom implementation, and Azure deployment engagements.
Source code available under NDA. Live demo using your sample data available on request.
Azure AI Engineer · LLMOps · MLOps · RAG · Multi-Agent Systems · Document Intelligence · FastAPI · LangGraph · Python