DRW

Quantitative Developer Intern

Parallelized and refactored several parts of a Python code base used to support Quantitative Researchers, increasing scalability and reducing latency by 30%
Found and fixed a high-latency-impact, low-reproducibility bug that occurred in production
Developed a framework for regression testing and reproducing production bugs in a semi-realistic setting
Supported other team’s C++ operations, and wrote small scripts to automate processes

Taught 10-725 Convex Optimization

Improve in-house computer vision algorithms for robust detection and locating of drones in various environments
Modify open-source code to fit client use cases, integrating it with various drone hardware and the existing code base
Work on improving UI/UX for current user-facing software
Develop drone algorithms for drone-vs-drone purposes (taking down another drone using a drone)
Conduct on-the-ground experiments to determine the efficacy of different approaches

Developed a containerized desktop and web based visualization software
Integrated various image and voice Deepfake detection models into the software, visualizing their gradient-weighted class activation mapping, enhancing the explainability of the team’s artificial intelligence products
Provided technical input on which direction the team should head in and how to improve their current solutions, which helped shape the team’s decisions

Thoroughly tested several state-of-the-art Automatic Speech Recognition (ASR) models to verify their suitability for deployment
Implemented a containerized end-to-end ASR pipeline with detailed documentation, allowing users to easily utilize the pipeline for their needs
Deployed the completed pipeline to distributed systems using ClearML, enabling fast distributed training on large internal datasets

Taught the following modules:

Nominated for teaching awards by 9 students