Sai Prabhakar
AI Researcher
ML Research Scientist
Tempus AI Inc.
Gen AI Team,
CA, USA
I graduated from Carnegie Mellon University with a Master of Science in Robotics (Thesis; MSR) degree, having completed my undergrad at the Department of Mechanical Engineering, Indian Institute of Technology Madras. My master’s thesis, conducted under the supervision of Prof. Manuela Veloso and Dr. Stephanie Rosenthal, focused on Explainable AI applied to Vision and Robotics. Prior to my current role at Tempus AI Inc., I worked at Abridge AI, where I spearheaded the development of conversation understanding capabilities for various healthcare actors, and led ML R&D for automatic doctor documentation.
My professional experience spans across both startup environments, where I built breakthrough ML products from scratch and scaled the company from 0 to 850 Million $, as well as large product and research-oriented companies. Additionally, I am a member of the International Consortium for Artificial Intelligence (AI) and Shared Decision Making special interest groups (SDM).
Academically, my interests lie in Health AI, Reinforcement Learning, Natural Language Understanding, and Machine Learning. Currently, at Tempus, I am delving into new avenues for Generative AI in Healthcare. Notable recent research includes fine-tuning LLMs, enhancing Summarization Models with Human Feedback and Edits, and devising new metrics for evaluating LLM-generated summaries and Speaker Diarization. My work has been published across top-tier conferences in fields such as LLMs, NLP, Speech, Computer Vision, Robotics, and Reinforcement Learning.
Throughout my career, I have worked extensively with popular Deep Learning frameworks, large datasets, Cloud platforms, and MLOps tools like Kubernetes and Pub/Sub.
affiliations
selected publications
- Improving Summarization with Human EditsIn Proceedings of main conference EMNLP, 2023 , 2023
- Generating more faithful and consistent SOAP notes using attribute-specific parametersIn Proceedings of the Machine Learning for Healthcare, MLHC, 2023 , 2023
- Medication Regimen Extraction From Medical ConversationsIn Proceedings of W3PHIAI of the 34th AAAI Conference on Artificial Intelligence, 2020 , 2019
- Learning End-to-end Multimodal Sensor Policies for Autonomous NavigationIn Conference on Robot Learning , 2017
- Dynamic generation and refinement of robot verbalizationIn 2016 25th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN) , 2016
- Verbalization: Narration of Autonomous Robot Experience.In IJCAI , 2016