• Facilitating the adoption of a new incident triaging workflow for rosbags using Foxglove Studio by assisting developers with bug fixes, addressing migration pain points, and providing ongoing support to ensure a smooth transition.

• Enabled quicker incident triaging by implementing cloud rosbag player and removing the need to download large rosbags locally.• Selected best solution by implementing and comparing multiple proof of concepts including third-party solutions like Foxglove Studio and custom-built alternatives using VNC or a VPN on an EC2 instance.• Enabled incident rosbag visualization in Foxglove by writing Typescript extensions to render custom perception and planning messages.• Facilitated migration from RVIZ to Foxglove by creating upload pipline for large rosbags using Python, AWS Lambda, and EC2.• Improved maintainability and ease of feature development on company-wide CLI tool by refactoring Python code, bash scripting, writing full coverage unit tests, and building CI/CD pipeline with Github Actions.

• Working on implementing a ROS behavior stack for modular robots.

• Developed scalable logging pipeline to process over 200TB per day of logs with Kubernetes and Vector.• Enabled usage across multiple clusters by implementing traffic routing and load balancing with Istio.• Implemented feature requests from multiple teams including rate limiting and log-based metrics.• Automated deployment using Helm and Buildkite.

• Led recitation classes, held office hours, designed exam questions and project test cases, and graded exams and projects for Honors Data Structures.• Student comments include, "He always went above and beyond to help us with the projects" and "Filled in other TAs' gaps in content/knowledge during discussion sections."

• Helped secure high-value contracts by rewriting Dockerfiles to eliminate critical security vulnerabilities and creating Go microservice to run dependency report generators.• Contributed to migration of 20+ microservices from self managed Kubernetes cluster to EKS by updating CI/CD pipelines and Helm deployment charts.• Implemented features and bug fixes to UI, Python SDK, and microservices of the Chariot MLOps platform.• Streamlined onboarding for 3 developers and 2 interns by writing new docs and giving platform walkthrough.

• Trained transfer learning neural networks in Python to track support of the fat acceptance movement on Twitter.• Implemented topic modeling and geoparsing to analyze tweets complaining about state unemployment systemsduring the pandemic using machine learning in Python.• Wrote scripts to collected and process tweets to analyze telehealth practices using Python and Twitter API.• Published 2 peer-reviewed papers, one as a joint lead-author, and qualified as an International Science and Engineering Fair Finalist.

• Raised autonomous IED defusal rate by 5x by implementing new state machine, and transforming and utilizing camera and LIDAR data in Python and ROS.• Improved robot reliability by repairing mechanical issues, designing new parts in CAD, and 3D printing them.• Increased team productivity by teaching and managing new interns.Position held for 6 and 8 weeks during summer 2020 and 2021 respectively.