Summary
Vivswan Shah is a machine learning systems engineer with nine years of experience building low-latency, distributed deep learning infrastructure across industry and academia. He holds a PhD focused on real-time distributed deep learning for heterogeneous GPU/CPU/NPU clusters and has led a research lab that deployed secure, high-throughput pipelines and optimized models for hybrid analog-digital accelerators achieving nanosecond-scale inference. His industry roles span applied science and research engineering at Amazon, Scale AI, and Microsoft AI, working on GenAI, agentic systems, LLMs, and MLOps for production at scale. He combines hands-on systems work (CI/CD, Terraform, Ansible, Docker) with model-level expertise (distillation, pre-/mid-training, evaluation) and experience integrating novel photonic and hardware accelerators into ML stacks. Based in the New York City area, he often bridges academic collaborations with industry partners including Google X and the DoD, bringing a rare mix of hardware-aware model optimization and production ML infrastructure.
9 years of coding experience
5 years of employment as a software developer
Bachelor's degree Computer Science and Physics, Bachelor's degree Computer Science and Physics at Illinois College
Doctor of Philosophy - PhD Machine Learning System, Doctor of Philosophy - PhD Machine Learning System at Carnegie Mellon University
Doctor of Philosophy - PhD Machine Learning System, Doctor of Philosophy - PhD Machine Learning System at University of Pittsburgh
English, Hindi