AI Software Engineer building full-stack AI products and modern data systems.

My work focuses on agentic systems, retrieval and knowledge platforms, AI evaluation, and Microsoft Fabric / Azure modernization. I care about systems that are grounded in reliable data, observable in production, and useful in real workflows.

View repositories · Open-source pull requests

• TransitionIQ — Workforce health transition workflows for employers, employees, consultants, and administrators. TypeScript React Express PostgreSQL

• FabricShift — Microsoft Fabric migration readiness, target mapping, reconciliation, lineage, and wave planning. Microsoft Fabric TypeScript React Python

• VANGUARD — Pull-request risk analysis, blast-radius mapping, CI diagnostics, and release-readiness gates. Python FastAPI Next.js Neo4j OpenSearch

• ContextMesh — Private knowledge platform with hybrid retrieval, graph-aware context, reranking, and grounded answers. Python FastAPI Qdrant Neo4j

• ARGUS — LLM red-team and model-risk evaluation for prompt, tool, privacy, citation, and guardrail failures. Python FastAPI LLM evaluations CI gates

• ResumeTailor — Local resume tailoring with reviewable changes and format-preserving DOCX output. React TypeScript FastAPI python-docx

• Agent systems: ORION, CALLSIGNAL
• Retrieval and knowledge: CITADEL, Evidence Graph RAG
• Data and analytics: HELIOS
• Learning: AI Engineer Roadmap 2026

• Pydantic #13435 — Fixed the pre-commit installation command in the contributor guide. Merged upstream.
• Vertex AI Python SDK — Agent evaluation, credential defaults, and prompt metadata: #6975, #6983, #6984
• vLLM — Parallel-sampling tests and ModelOpt documentation: #48062, #48063
• Semantic Kernel #14145 — Preserved explicit None defaults in Python function metadata.
• Altair #3300 — Added support for deeply nested schema types.
• ComparIA #596 — Repaired and documented the dataset export workflow.
• Kubeflow Pipelines #13703 — Routed SDK formatting through project make targets.

• Languages: Python, TypeScript, JavaScript, SQL, Java, C#
• AI: Azure OpenAI, LangGraph, RAG, LLM evaluation, MCP
• Data: Microsoft Fabric, PostgreSQL, Neo4j, Qdrant, OpenSearch, DuckDB
• Platform: Azure, AWS, GCP, Docker, Kubernetes, Terraform, GitHub Actions

• Ground answers in trusted data and explicit evidence.
• Keep workflows observable, reviewable, and recoverable.
• Evaluate behavior before treating a model as reliable.
• Build complete products, not isolated demos.