Summary
Arvind Somasundaram is a Machine Learning Engineer with eight years’ experience building computer vision, generative AI and statistical ML systems, currently developing Catalog Intelligence models at Meta. He has a strong research-to-production background—from doctoral work on differentiable physics and cloth manipulation transferred to real-world robots to industry projects that extract human measurements and reconstruct bodies from video. His career blends academic rigor from TUM and ETH Zürich with product-focused roles at startups and Meta, including work on lip-synced multilingual video generation and large-scale NLP tools for StudySmarter. Comfortable across pose estimation, 3D human mesh fitting, and end-to-end ML systems, he pairs deep technical expertise with practical deployment experience in the San Francisco Bay Area. An under-the-radar strength is his track record of converting complex research (e.g., implicit N-D representations and optimization layers) into robust, deployable solutions.
8 years of coding experience
2 years of employment as a software developer
Master of Science - MS, Informatics, 1.6 / 1.0, Master of Science - MS, Informatics, 1.6 / 1.0 at Technical University Munich
Bachelor of Engineering (Honors) - BE (Hons), Computer Science, 8.35 / 10.0, Bachelor of Engineering (Honors) - BE (Hons), Computer Science, 8.35 / 10.0 at Birla Institute Of Technology and Science, Pilani Dubai
High School Diploma, High School Diploma at National Public School