- Implemented C++ based tensor operators to enhance neural network model inference and optimize real-time image processing in Advanced Driver-Assistance System (ADAS) models- Developed a Bash script that drove the creation of a specified model (like ResNet-18) and compared the outputs of the Python and C++ versions against a set threshold, thus automating the testing process- Proactively enhanced the Bash script to support testing of both C++ based operators and models, achieving an 85% reduction in test execution time- Integrated automated end-to-end testing using CI/CD, improving system efficiency by 85% through continuous bug detection and codebase improvements- Developed a comprehensive project documentation and environment setup instructions, reducing setup time from 1.5-2 weeks to just 2 days- Refactored existing code to make codebase design more modular and function-driven- Gained hands-on experience for end-to-end software development life cycle in a complex technologyenvironment

- Spearheaded the implementation of a comprehensive mentorship program, connecting 50+ studentswith industry mentors, which helped enhance networking opportunities- Conceptualized and launched an interactive calendar of events for the upcoming semester, whichincreased student engagement by 17%

- Assisted in weekly tutorial sessions to guide 15+ students through questions based on discretemathematics and proofs- Conducted 1-on-1 office hours to help students develop well-constructed proofs and deepen theirunderstanding of fundamental theoretical concepts

- Resolved critical issues in the risk model that was unable to detect some of the factors responsiblefor a major financial loss for the bank in 2021- Implemented API integration to efficiently retrieve client account information from the database- Computed the average correlation and covariance based on the aforementioned data, which weresubsequently used in calculating risk factors