MrJuneJune

Hands-on engineering leader building AI agent execution platforms and production software at Microsoft, Meta, Google, and growth-stage companies. Foundational engineer for Microsoft's AIX Harness, spanning agent orchestration, context management, and multi-model execution. Led zero-to-one delivery from platform architecture to public product launches.

• Served as one of the lead engineers for Copilot Tasks from architecture through research-preview launch, focusing on Word, Excel, and PowerPoint generation; web-app generation and sharing; browser actions; and AI-powered mobile-browser experiences.
• Designed and built AIX Harness as a foundational engineer and technical leader: the execution and control plane for the Copilot SuperApp. Standardized agent orchestration, context management, and multi-model execution, supporting 10K pull requests per month across dependent teams.
• Created the repository and led end-to-end engineering for the Code and Autopilot tabs, two flagship SuperApp surfaces introduced publicly by Satya Nadella at Microsoft Build 2026.
• Set architecture and coordinated cross-team delivery across Copilot Tasks, AIX Harness, and the SuperApp codebase, taking multiple zero-to-one systems from definition to launch in nine months.

• Built Racer Pacer, an internal automated code-review bot powered by Llama models on Meta's existing Racer infrastructure; reviewed around 1K pull requests in one month.
• Initiated Channel Value Rule work for the 16% of ad traffic with both app and web destinations, targeting more accurate value attribution and advertiser ROI.
• Built full-stack advertising features with React, Hack, and GraphQL; partnered with data science to design A/B tests and analyze the revenue impact of ad destinations.
• Proposed and implemented improvements to internal testing infrastructure, reducing test time by 50%.

• Established hermetic Bazel builds and shared testing and code-organization standards for the company's TypeScript and JavaScript codebase.
• Led five engineers building client-facing GraphQL APIs with Apollo and AppSync, supporting over 2,000 requests per second with request-driven autoscaling.
• Reduced GraphQL response times by up to 85% through database schema and SQL improvements, eliminating N+1 queries and adding missing indexes.
• Designed pub/sub, caching, and media-processing infrastructure and developed backend CI/CD pipelines.

• Built and maintained Apps Script features across Google Workspace, including Gmail, Sheets, and Docs.
• Improved response and render times for Apps Script hover-card components.
• Conducted user testing and incorporated feedback into feature delivery and usability improvements.

• Maintained AWS Amplify and React applications delivering at-home COVID-19 test kits for enterprise customers, including the Brooklyn Nets and Tinder.
• Automated assessment and refunds for unused test kits, saving up to $200,000.
• Built Rails tooling for end-to-end QA workflows and bulk customer refunds, reducing support tickets by 50%.
• Added record-level audit trails with PaperTrail to trace production changes to the responsible process.

• Built REST APIs in Django, Ruby on Rails, and Flask; automated API documentation with Swagger.
• Built Rake tasks to migrate production data into updated schemas for client applications.
• Maintained staging and production services on Heroku and AWS and debugged PostgreSQL production issues.
• Built Python and Selenium test automation that eliminated 80% of manual QA work.

• Developed a Shapley-value regression model to explain feature importance using scikit-learn, pandas, and NumPy.
• Built an Amazon S3 image-upload integration with boto3.
• Built Flask REST APIs with pytest tests and automated API documentation using flask-restplus and Swagger.
• Built interactive D3.js visualizations in Vue.js backed by Flask APIs.