Aditya Kumar Email & Phone Number

Fremont, CA, US

- Explore boolean satisfiability (SAT) problems with up to 250,000 variables; study the concepts of NP-completeness- Design experiments to test the efficacy of different solving methods for SAT problems- Develop hardware-accelerated ML heuristics inspired by state-of-the-art reinforcement learning algorithms

Los Angeles, California, United States

- Designed algorithms and wrote programs for reading transistor netlists and detecting redundant aspects of these circuits- Implemented C++ data structures to store large circuits and simulate the resulting truth tables- Presented findings in yearly symposiums with the other CURVE (Center for Undergraduate Research in Viterbi Engineering) research fellows

Los Angeles, California, United States

- Engage with an interactive curriculum to learn basics of AI, ML models/metrics, and neural network architectures- Develop personal notebooks to parse datasets, preprocess data, and practice creating different machine learning models- Liaise with others in curriculum group to develop understanding and promote discussions on AI for social good- Collaborate.

Sunnyvale, California, United States

- Refactored and extended evaluation pipeline for Workspace query understanding with added tests and a scalable structure- Developed output diff reports with custom configurations and metrics per error type, streamlining LLM error analysis- Built a prompt engineering pipeline and conducted experiments to optimize prompt size while preserving accuracy.

Los Angeles, California, United States

- Provided conceptual help and debugging assistance for students in CSCI103 (C++ Programming) during office hours- Led lab sessions to teach concepts such as compilation, memory allocation, Linux, and object-oriented design

San Jose, California, United States

- Simulate over 10,000,000 transactions for scale testing of ZPA ML models to create access policies for users and apps- Examine new heuristics to efficiently form user and app clusters using collaborative filtering recommendations- Create integration tests hosted on Kubeflow pipelines to confirm validity of policies and user/app groupings- Implement a.

San Leandro, California, United States

- Created a Picterra machine learning model to identify homeless tents given satellite imaging of a city in order to relocate the homeless population- Initiated discussions with Los Angeles homeless shelters regarding the implementation of our model

- Organized a team to represent our school in competitive tournaments- Created lesson plans for students new to chess

- Participated in over 150 nationally-recognized tournaments- Ranked in the 90th percentile nationally

Bachelor Of Science - Bs, Computer Science, 3.97 Out Of 4.0

Activities and Societies: CURVE Undergraduate Research Assistant | Viterbi Impact Program member | Center for AI in Society's Student.

4.0 Out Of 4.0

Activities and Societies: Chess Club (co-vice president) | Astrophysics Club (activities commissioner) | Computer Science First Club.