– Modeled risk-adjusted PnL-maximizing strategies in Python, analyzing and processing over 1.2 billion data points.– Designed an object-oriented testing framework for quick iteration over 10+ strategies, and scalable workloads.– Plotted 100+ graphs over ∼20 simulations, to explain to team members the best strategy per situation.

– Assisted in 6.1910, a class of 230 students, to help identify and resolve bugs in laboratory assignments, and clarify understanding of concepts.– Labs consisted of introduction to RISC-V assembly, building a processor from scratch with a hardware description language, and performance engineering a C snippet.– Debugged successfully 100% of student labs while also explaining what went wrong with whiteboarding.

– Optimized SQL SELECT queries for compressed NoSQL JSON array elements in relational databases.– Eliminated unnecessary and expensive decompression for a potential 7x speed improvement.– Scanned through tens of thousands of lines of C++ code in GDB over 10+ files to systematize code modifications.– Completed implementation for ∼75% of the various edge cases of the database table schema

– Contributed to the lab's project, ATL, a verified programming language for specifying tensor operation schedules via semantic-preserving schedule rewrites proven as theorems in the Coq proof assistant.– Refactored code to create a general-purpose proof tactic that allows any number of rewrites of the Im2Col tensor optimization to vectorize image convolution operations using Coq's pattern-matching language in Ltac.

– Soldered 6 PCBs with a team of 6, for safe power I/O of 50V LiPo batteries and emergency power isolation.– Reported results asynchronously to two overseas PCB designers detailing 10+ testing procedures and voltagemeasurements, complete with wiring diagrams, pictures of the circuits, and Slack feedback messages.– Designers were eventually convinced that the issues stemmed from the PCBs, not from faulty soldering or probing.– Achieved a working robot with emergency power isolation, presented in the 2022 NASA BIG Idea Challenge.