🤖 Tony (Minxing) Pan

Perception Data - curate data for foundation model data engine

Simulation Infrastructure - largescale distributed system with ML accelerators

Teaching assistant for EECS 442 Computer Vision • Teach a recorded hybrid discussion section for 80+ in person and Zoom students • Design weekly problem sets and write solutions and rubrics for them • Hold office hours and answer Piazza questions for ~250 students

• Conducted experiment to determine the computation and latency savings of databases • Wrote 10k+ lines of C++ and Python code and tests for database migration and worked on design for the migration • Safely handled migration in real time server that processes massive traffic with millions of queries per second • Designed, implemented, validated, and launched the tracking of a new type of user interaction to database • Implemented solution for a customer issue with revenue impact on Google and securely monitored the changes • Independently refactored several complex components for different teams to improve their database migration process

Intelligent robotics research with the SOAR Cognitive Architecture under EECS professor John Laird.

• Collaborated with two other interns to build a full stack web app on the Google Cloud Platform that recommends nonprofits to users • Developed Machine Learning workflows to perform text similarity analysis and categorization with TensorFlow in Python • Designed and created a Cloud SQL for MySQL relational database for querying and ranking • Developed a recommender system for nonprofits using SVD for collaborative filtering • Wrote programs and worked with another team within Google to scrape and process 30,000+ nonprofits from the Internet I also worked on a side project at Google during my internship: • Created very memory efficient program to calculate user height above sea level for any land mass from GPS altitude readings, using < 20 KB of memory with < 1 m accuracy (original size ~ 1 GB) • Performed statistical and graphical analysis on performance of different compression, resizing, and masking parameters

Participant in the 50th Dr. Bessie F. Lawrence International Summer Science Institute (ISSI). Conducted research project "Making the Invisible Visible: Optimizing Laser Speckle Imaging Using Machine Learning and Targeted Recoloring" with Prof. Vyacheslav Kalchenko