Engineering leader with 12+ years of experience building large-scale developer platforms and AI-enabled developer experiences. Currently leading AI systems that connect frontier model performance to real-world developer workflows — owning model evaluation, agent runtime infrastructure (terminal and browser), and feedback loops used by internal and external model providers to improve coding model performance. Lead engineering across multiple Microsoft developer products including Visual Studio Code and Copilot CLI, and continue to lead the Python developer ecosystem used by ~14M developers monthly. Known for operating at the intersection of model development and developer tooling, with evaluation systems recognized by model providers as among the most advanced real-world coding benchmarks.
AI Systems & Platform Impact
Contribute to developer platforms serving ~45M developers globally.
Lead the Python ecosystem used by ~14M monthly developers.
Built infrastructure powering debugging, evaluation, and AI-driven workflows across Microsoft developer tools.
Designed evaluation systems recognized by model providers as among the most advanced real-world coding benchmarks.
Professional Experience
Principal Engineering Manager — MicrosoftSept 2024 – Present
AI Systems for Coding Agents · Model Evaluation · Agent Runtime Infrastructure
Lead engineering for AI systems powering coding agents across Microsoft developer platforms — spanning model evaluation, runtime environments, and feedback loops used to improve model performance.
Lead an engineering team of 8 (including tech leads) across AI evaluation systems, agent runtime infrastructure, and developer tooling platforms.
Own AI model evaluation and feedback systems used across multiple products including Visual Studio Code and Copilot CLI, benchmarking 25+ frontier models.
Designed and implemented advanced evaluation frameworks consisting of: ~60 high-fidelity benchmark scenarios modeling real developer workflows; 5 benchmark suites executed across 70–100 cases per suite; repeated evaluation runs enabling statistically robust comparisons.
Provide structured evaluation data used for model selection and product integration decisions, performance optimization (latency, token efficiency, tool usage), and model training and pre-alpha development with internal and external providers.
Partner directly with model teams (internal and external) to improve coding model performance across code quality, tool-calling behavior, trajectory efficiency, and real-world task completion rates.
Lead development of agent runtime infrastructure, including terminal systems enabling code execution workflows, browser systems enabling web interaction, data extraction, and vision-based workflows, and tooling used by AI agents to operate across environments.
Drove adoption of evaluation and runtime systems beyond VS Code into Copilot CLI, expanding platform impact across products.
Act as a cross-organization technical leader advising multiple teams on AI system design, model integration, and evaluation strategy.
Define roadmap and technical direction for AI-powered developer systems, influencing both product capabilities and model development priorities.
Continue to lead the Python ecosystem in VS Code including extensions, debugging infrastructure, and developer workflows used by ~14M monthly developers.
Principal Software Engineer — MicrosoftFeb 2024 – Sept 2024
Technical leader for Python developer tooling across the VS Code ecosystem.
Technical leader responsible for Python developer tooling across the VS Code ecosystem.
Led architecture and development of major debugging and language tooling improvements serving millions of developers.
Directed cross-team engineering efforts improving Python development workflows and debugging reliability.
Senior Software Engineer — MicrosoftFeb 2017 – Feb 2024
Core developer of VS Code Python ecosystem; creator of debugpy (2.4K+ stars); LSP tooling architect.
Core developer of the Python developer ecosystem in VS Code including linting, formatting, language analysis, testing, and debugging tooling.
Designed and built debugpy, the modern Python debugger used across developer tools, cloud environments, and notebook platforms.
debugpy has 2.4K+ GitHub stars and powers debugging across VS Code, Visual Studio, Jupyter environments, and multiple cloud platforms.
Designed and developed tooling for the Language Server Protocol including libraries and protocol code generation systems.
Built multiple VS Code extensions used by millions of developers globally.
Software Development Engineer II — MicrosoftFeb 2015 – Feb 2017
UWP platform APIs; SQLite integration; Windows CVE resolution; remote R tooling for Visual Studio.
Worked on platform APIs for Universal Windows Platform.
Integrated SQLite as a Windows platform component reducing application package sizes by ~1.2MB across the ecosystem.
Resolved platform CVEs reducing Windows attack surface.
Developed remote R tooling enabling seamless local and remote data science workflows within Visual Studio.
Software Development Engineer — MicrosoftAug 2012 – Jan 2015
Async causality tracking; Windows app componentization; per-app registry hive architecture.
Developed asynchronous causality tracking infrastructure enabling IDEs to display complete async call stacks across language boundaries.
Designed Windows application componentization mechanisms enabling applications to expose reusable APIs to other apps.
Built per-application registry hive architecture reducing runtime memory overhead for Windows applications.
Open Source & Platform Impact
VS Code Python Ecosystem — Core maintainer of the Python extension platform serving ~14M monthly developers.
debugpy — Python debugger used across developer tools, cloud platforms, and notebook environments (2.4K+ GitHub stars).
lsprotocol — Language Server Protocol tooling and code generation library.