About
Samuel Herman
I am a distinguished engineer and researcher working on retrieval science, language models, and visual intelligence, and on the end-to-end capabilities that make them run at scale and in the physical world. I have led engineering organizations of more than 75 people at AWS, Oracle Cloud Infrastructure, and DataStax, building search, SIEM, and observability systems that process petabytes of data a day in real time.
I am a member of the OpenSearch Leadership Committee and hold patents in distributed systems and storage. My research predicts how approximate indexes behave before they are built, prices retrieval at production scale, and extends the same measurements to computer vision and perception at the edge.
For technical leadership, advisory, or research collaboration, reach me on LinkedIn.
Short bio for organizers
Samuel Herman is a distinguished engineer and researcher working on retrieval science, language models, and visual intelligence, and a member of the OpenSearch Leadership Committee. He has led engineering organizations at AWS, Oracle Cloud Infrastructure, and DataStax building systems at petabyte scale, and holds patents in distributed systems and storage. His research, including the arXiv paper Closed Forms and Synthetic Twins, predicts the behavior and cost of approximate retrieval from embedding geometry, and extends the same methods to computer vision and AI in the physical world.
Elsewhere