Summary
Arsh Zahed is a software engineer specializing in deep learning for speech processing, reinforcement learning, and computer music, with nine years of experience across industry and academia. He has driven research-to-production work at TikTok, NVIDIA, Google, and Together AI, and currently focuses on PyTorch compilation at Meta while teaching deep generative models at Stanford. His background spans speech synthesis, voice conversion, audio enhancement, and model-based RL, blending theoretical research with practical deployment. As an undergrad leader at UC Berkeley’s Launchpad and a long-serving officer, he has a track record of mentoring and scaling student ML projects into real outcomes. Notably, he combines production-focused engineering (Riva, deployment of foundation models) with hands-on research experience in conversational AI and generative audio. Based in San Jose, he leverages cross-disciplinary training from UC Berkeley and Stanford to move novel audio and ML ideas into robust systems.
9 years of coding experience
6 years of employment as a software developer
Bachelor of Science (B.S.) Electrical Engineering and Computer Science Minor in Mathematics, Bachelor of Science (B.S.) Electrical Engineering and Computer Science Minor in Mathematics at UC Berkeley College of Engineering
Computer Science, Computer Science at Stanford University