Rumit Desai

Background

About Rumit Desai

Currently working on building software stack for MTIA (Meta’s Training and Inference accelerator)(https://ai.meta.com/blog/next-generation-meta-training-inference-accelerator-AI-MTIA/). I am currently leading the inference runtime area to productionize the SW stack for revenue-critical models.Before joining MTIA-SW I worked with Meta’s AI Infra Inference platform: Designing and building highly reliable inference-as-a-service platform for hosting/serving Meta’s mission critical models.My software engineering experience dates back to my first job in 2016 at Altair Engineering where I was part of back-end framework(data engine) team, solving problems related to efficient FEA data handling. In 2020 I joined Zenuity(later rebranded as "Arriver" which was acquired by Veoneer Systems Software which was furthter acquired by Qualcomm) as a software engineer on Vehicle controls team working closely with control theory algorithm engineers to deploy code efficiently.I completed my M.S. degree from State University Of New York-Stony Brook. During my time at SUNY, I also did a summer internship at Zebra Technologies in computer vision domain, where I worked on a research project for their healthcare vertical. Internship experience inspired me to pursue my research thesis in human computer interaction where we developed software tool to capture human motion data using low cost depth sensor.