Bits, Bytes & Vision

Exploring the intersection of high performance computing, modern C++, computer vision, and storage systems. Deep technical insights into building performant software systems, optimizing databases, and implementing efficient computer vision algorithms.

---

Project maintained by sam-herman Hosted on GitHub Pages — Theme by mattgraham

Samuel Herman

Strategic technology insights with in-depth analysis of emerging trends in AI infrastructure, Search, and distributed systems.

Distinguished Engineer & Technical Visionary

Distinguished engineer with 10+ years building Search, SIEM, Observability, and AI technologies at scale.

I’ve led teams of 75+ engineers at AWS, Oracle Cloud Infrastructure, and DataStax, shipping products that process petabytes of data daily in real-time. Author of patents in distributed systems and storage technologies. Currently serving as OpenSearch Leadership Committee Member, driving technical strategy for the OpenSearch project.

Current Focus

• Computer Vision
• Storage & Vector Databases
• Edge Inference
• Large-scale AI Infrastructure

---

Recent Publications & Talks

Teaching Embedding Models New Words: A Deep Dive into Domain Adaptation

Embedding Adaptation for Domain-Specific Search

Breaking the Single-Thread Bottleneck

Concurrent Vector Graph Construction in OpenSearch

The Reality of Vector Quantization

Does Product Quantization Deliver on Its Promises?

Are GPUs Good For Vector Search?

GPU Performance Analysis for KNN Search

Conference Talks

OpenSearchCon 2025 - Vector Search Beyond The Hype - OpenSearch and Vector Databases OpenSearchCon 2024 - Lucene And Beyond - Core Storage Extension In OpenSearch

OpenSearchCon 2023 - Journey To 750 Node Clusters

---

Interested in technical leadership, consulting, or collaboration? Connect on LinkedIn