Noah Hein is a senior product marketing engineer at Tiger Data, a former founding engineer at Smol AI, and an educator who co-created Latent Space University. His work spans developer education, AI-powered applications, and the database infrastructure needed to operate them safely.
A self-taught programmer who previously managed fast-food restaurants, Hein moved into developer advocacy at QuickNode and later worked at Phantom. He subsequently helped develop Latent Space University’s introductory AI curriculum, leading its first public workshop at the 2023 AI Engineer Summit. Participants built a Telegram application incorporating conversational memory, document retrieval, generated code, image generation, and speech transcription.
At Smol AI, Hein worked on Smol Talk, the platform behind AI News. He later taught agentic research workflows using LlamaIndex, covering coordinated agents, streaming, tracing, and iterative review. At Tiger Data, he writes about PostgreSQL and time-series infrastructure, including high-volume deletion in TimescaleDB and zero-copy database forks.
- Build products to understand models. Hein teaches AI engineering through working applications: API calls, conversation state, debugging, interfaces, and progressively integrated language, image, and speech capabilities.
- Retrieval quality depends on data quality. His AI Engineering 101 workshop demonstrates how token limits, embeddings, chunk boundaries, overlapping passages, and incomplete documents affect grounded answers. He recommends starting with local arrays before introducing specialized vector databases when genuine performance demands justify them.
- Give coding agents safe operating environments. Hein uses generated code and conventional software to compensate for unreliable model reasoning, including mathematical operations. His work on zero-copy database forks extends that principle to isolated, production-shaped environments for testing agent-generated database changes.
- Operational data makes physical AI deployable. His writing on robotics and autonomous systems emphasizes telemetry, safety constraints, regulatory requirements, edge cases, and historical records alongside model performance.
- Responsive interfaces are essential infrastructure. Hein treats streaming and visible progress indicators as fundamental when transcription, image generation, or other model operations introduce noticeable delays.
