About

Aditya Bhagwat is an AI Framework Engineer at Modular, with a passion for accelerating deep learning on heterogeneous hardware. He has a strong background in distributed systems, design, and development of large-scale applications, as well as high-performance libraries. Aditya is experienced in training huge deep-learning models using Pytorch, and has intermediate C++ skills with an interest in profiling performance. He is skilled in optimizing performance by leveraging parallelism and concurrency, with expertise in MPI, OpenMP, and CUDA. Aditya is passionate about low-level systems programming in C/C++ and Rust. Aditya's language proficiency includes C++, Python, Rust, Java, Golang, Ocaml, CUDA, and Racket. He is also familiar with various technologies such as Pytorch, Apache Spark, Airflow, Docker, Kubernetes, Spinnaker, Jenkins, AWS/GCP, Neo4J, Redis, MySQL, SpringBoot, Django, and MongoDB. Prior to joining Modular, Aditya worked as a Research Engineer Intern at Snapchat, developing ML infrastructure for training Graph Neural Networks at scale. He also served as a Research Assistant at Carnegie Mellon University, working under Prof. Eric Xing to investigate methods for accelerating large-scale model training using parameter reduction. Aditya developed solutions to reduce the communication overhead in the distributed training of large-scale models, leveraging matrix factorization and sparsification methods. Aditya holds a Master of Science in Computer Science and Machine Learning from Carnegie Mellon University, as well as a Master of Technology in Computer Science and a Bachelor of Technology in Computer Science from the Indian Institute of Technology, Kharagpur. He also completed his primary, middle, and high school education at Bhavan's Bhagwandas Purohit Vidya Mandir in Nagpur. Aditya's technical skills and relevant experience make him a valuable addition to any team in the AI and machine learning space.