Software Engineer II at Microsoft (Azure Core Insights). I build and operate large-scale backend and data systems: services, pipelines, and the infrastructure that keeps them reliable at fleet scale.
- 🔭 At work: petabyte-scale data pipelines and ML systems that predict and prevent hardware failures across Azure's global fleet, including the ingestion, orchestration, and deployment systems behind them.
- 🧠 Focus: backend services, distributed data systems, reliability, and API/agentic tooling.
- 📍 Seattle, WA
- 📝 Co-author, AFPD: System for Prediction, Mitigation and Remediation of Node Interruptions in Cloud Infrastructure, IEEE Cloud Summit Best Paper 2025 (Microsoft blog)
Pac-Mapper is for runners and
cyclists trying to cover every street in their city. It syncs Strava activities,
map-matches GPS traces against OpenStreetMap streets using PostGIS, visualizes
street-level coverage with Leaflet, and suggests OSRM routes to explore an entire city,
like Pac-Man eating dots on a map. Built with FastAPI + React, deployed on Azure
(App Service, ACR, PostgreSQL Flexible Server).
Live demo · Code
Python · FastAPI · PostGIS · React · TypeScript · Leaflet · OSRM · Azure
mcp-data-detective is an MCP
server that turns any LLM into an interactive data-engineering copilot: connect,
query, profile, and monitor SQLite/Parquet/CSV sources through natural conversation.
Validated against the 20.9M-row NYC TLC dataset.
Python · MCP · DuckDB · FastMCP
Backend & data: Python · FastAPI · SQL · KQL · PostgreSQL / PostGIS · SQLAlchemy · REST APIs · Docker
Cloud & infra: Azure (App Service, ACR, Functions) · Azure Data Explorer (Kusto) · Databricks · Spark
ML & tooling: PyTorch · MLflow · Hyperopt · MCP · RAG · DuckDB
Also work in: C# · TypeScript · Scala