Bryan Kwan Email & Phone Number

I engineer Vision Transformer architectures for visual relation detection and scene graph generation.

I design membership inference attacks on explainable ML models.

Los Angeles, California, United States

I research mechanistic interpretability of LLMs, looking into transformer circuits, induction heads, and the role of the residual stream. I also took a fellowship course on AI Safety issues and how to build a transformer language model.

Los Angeles, California, United States

I host and organize alumni UPE events for anybody at UCLA to come and learn more about life after college. Recently, I led a private dinner mixer with over 20 students and alumni attending.

Los Angeles County, California, United States

I am a projects officer, designing AI projects that students learn from and helping them implement models such as transformers. I engaged with deep learning concepts, applications of neural networks, and machine learning theory through AI workshops. After learning about these concepts, I created two hands-on deep learning models (bird call identification, sentiment analysis) using PyTorch.

Los Angeles, California, United States

I utilized Pandas, Matplotlib, and Seaborn to derive insights from serial killer datasets and create data visuatlizations. I collaborated with others to write on blog based on our analysis and presented our conclusions on DataRes Demo Day.

San Mateo, California, United States

I constructed and refactored a Python algorithm utilizing the Nomad API to upload PBS videos on to AWS cloud from the MongoDB database. During the process, I remodeled the video transfer system using Python object-oriented programming to make the code more effective, clear, and concise. The project helped enhance my skills on data transfer, database management, APIs, JSON file manipulation, and MongoDB via project/online courses. In order to complete the project, I leveraged Postman to test and communicate with the Nomad API involved with the video transfer from PBS to Roku

Berkeley, California, United States

Used Google Dialogflow to create natural language processing chatbots that answer people’s medical questions, enabling easier and quicker access to medical resources. Attained AI knowledge on Natural Language Processing, through online courses such as Google Qwiklabs.

San Mateo, California, United States

Led and created interactive freshmen orientations to instill active listening, empathy, and communication through storytelling, activities, and more

Los Angeles, California, United States

I engaged with and presented hands-on research using bioengineering instruments like spectrometers, chromatographs, and Instron machine in an UCLA lab run by PhDs and a UCLA Professor.

Bachelor Of Science - Bs, Computer Science, Sophomore

Activities and Societies: ACM Artificial Intelligence Took CS31 (Intro to C++), MATH170E (Introduction to Proability and Statistics), CS32.

High School Diploma