Experience

Architected a centralized GraphQL API with 25+ endpoints and server-side caching, unifying 3 internal data sources. Designed an A/B eval framework orchestrating 74 AI agent runs across 37 production replays, lowering root-cause time by 40%. Launched 5 bespoke debugging runbooks with 30+ real-time components, enabling faster severe incident (SEV) diagnosis.

June - August 2026

Engineered a consistent disaster recovery framework for 80,000+ distributed containers powering key infrastructure. Deployed Rust-based automated recovery for 400+ key services, enabling seamless support within the Meta family of apps. Enabled 15+ customers to modify and actively monitor their recovery packaging with Unidash GUIs and a Superconsole TUI.

May - August 2025

Designed a custom version of the SZ3 framework to compress terabytes of molecular charge data by 100x. Developed a Graph Neural Network for compressing volumetric data in under 1 second, significantly reducing dimensionality. Automated the aggregation of statistics from 5 compression algorithms, enabling detailed analysis of impacts on charge density.

September 2023 - May 2025

Completed a 12-week program with guidance from Meta engineers, gaining hands-on training and insights from software talks. Secured admission into highly competitive mentorship program, placing in the top 2% of applicants. Developed and deployed an internal networking template for engineers utilizing a VPS, MySQL, NGINX, Docker, and Flask.

June - September 2024

Analyzed working processes of 150+ trade companies and identified points of weakness, improving operational efficiency. Conducted market research on AI-based document segmentation models to save clients up to $500k a year.

September - December 2024

Utilized native multithreaded support in ROS2 to 4x path planning computations with LIDAR laser scanner. Led a team of developers in streamlining the migration of foundational software for Project Sprite from ROS to ROS2, enhancing system performance and reliability. Crafted unit and integration testing (90% code coverage) using Python and C++ for obstacle detection algorithms.

August 2023 - December 2024

Designed and implemented over 10 functionalities for the robot, including grasping and a bipedal walking algorithm. Enhanced Zeus2Q's capabilities by integrating the NVIDIA Riva SDK for real-time conversational AI using industry-leading large language models (LLMs). Developed technical documentation for Zeus2Q's Python-based interfacing API through documentation videos.

August - December 2022