Yash Mathur
Capstone Ambassador
Hey! My name is Yash, and I am currently pursuing my Master’s in Electrical Engineering and Computer Science at UC Berkeley, specializing in Data Science and Machine Learning. I previously completed my Bachelor’s at Berkeley, earning a double major in Computer Science and Data Science, and I’m excited to be a Double Bear!
Before returning for graduate studies, I worked as a full-stack software engineer at Newfront (YC W18), where I built in-house payment processing systems for insurance and developed a suite of LLM-powered legal tech products for contract review and certificate compliance. Through this work, I gained extensive experience designing intelligence pipelines and leveraging fine-tuned, domain-specific LLMs to extract, compare, and validate complex policy and contractual data.
Prior to this, I worked at NVIDIA with the GeForce NOW NLP team to deploy a BERT-based language model that categorized user-reported issues and integrated these insights into large-scale performance monitoring systems. Earlier in my career, I conducted research at the Rolls-Royce@NTU Corporate Research Lab in Singapore, developing ML pipelines for image classification and fault detection in jet-engine fan blades using x-ray imaging data.
I’m deeply interested in the intersection of language models, AI systems, and autonomous vehicles. I am currently pursuing my graduate capstone research project, affiliated with Berkeley AI Research (BAIR) Lab, where my team is investigating how hardware-induced computational delays impact control and perception in various autonomous driving scenarios. Additionally, I am also working with the Berkeley Center for Responsible, Decentralized Intelligence on researching multi-agent AI misalignment and oversight evaluation frameworks.