Noah Hein

Noah Hein is an AI engineering educator who led the first public run-through of Latent Space University material and co-created an AI fundamentals course with Latent Space for software engineers moving into AI engineering. Maven credits him and Shawn “Swyx” Wang for Level Up From Software Engineer to AI Engineer. His work also includes building AI software: the instructor biography accompanying his September 2024 roadmap lesson described him as Smol AI’s founding engineer and the solo engineer building Smol Talk, the platform powering AI News.[1][2][3]

In What Comes After the Pull Request, published March 16, 2026, Hein argues that faster code generation exposes a slower process of review, comprehension, verification and shipping. His central claim is: “When agents can produce large volumes of code quickly, the limiting factor stops being authorship and becomes evaluation.”[5][6]

He examines three responses: AI-assisted review that feeds findings back to coding agents; moving human judgment upstream into specifications and guardrails, a proposal he attributes to Ankit Jain; and platforms designed around agents as participants in software delivery. Hein argues that engineers increasingly contribute intent, constraints and feedback systems rather than inspecting every generated line. He also qualifies the promise of specification-driven development: a specification precise enough for machine execution begins to resemble a programming language, leaving an unresolved bridge between human intent and executable instructions.[5][6]

Hein treats these supporting systems as potentially temporary: stronger models may absorb capabilities that currently require elaborate orchestration or review tools. His recommendation is to solve present bottlenecks while keeping those layers thin and replaceable. As he puts it, “The layer may be temporary. The learning is not.”[5][6]

Before that writing, Hein’s teaching focused on helping software engineers build with AI systems. At AI Engineer Summit 2023, he led the introductory workshop that first brought Latent Space University material to a public audience. The publisher’s November 13, 2023 account describes five small projects covering GPT-3 API basics, prompt tooling and memory, GPT-4 code generation, image generation and speech-to-text with Whisper. Hein describes the course he co-created with Latent Space as a three-week introduction aimed at engineers transitioning into AI engineering.[1][2]

For July 16, 2025, Maven advertised Build Your Own “Deep Research” with LlamaIndex, jointly taught by Hein and Laurie Voss. Its planned workflow chained Research → Write → Review agents through shared context. The syllabus also covered streaming outputs and tool calls, tracing, and review cycles in which agents critique and improve their work.[4]

Hein’s retrospective account on his free course site places restaurant work before his engineering career: he recalls frying chicken for $14 an hour without a computer-science degree, then teaching himself to code and entering production AI engineering. He connects his decision to offer the course free to that experience: “I'm giving away the course for free because I remember what it's like to feel stuck while everyone else seems to be shipping.”[7]

The five-day course offers projects in Discord-based retrieval-augmented generation, a speech-to-image pipeline, automated evaluation, and an agent integrating Google Calendar and Gmail. It assumes Python and web-development experience and focuses on engineering systems with LLMs. Its stated qualification is practical: the applications are foundations requiring further work, rather than turnkey production products.[7]