The Agentic Pipeline
Build a real autonomous AI system, end to end
Watch: what you'll build and how the course works
Build a real agentic system, end to end
Learn agentic engineering by building a real AI system from scratch. You'll create an autonomous Python pipeline that turns a daily news stream into a polished startup-idea newsletter — complete with a header image and a podcast episode — running on your own infrastructure. Not a demo that works once on your laptop: a real system you build incrementally and understand completely, because you wrote every part of it.
You'll build it with Claude Code (or your agent of choice), prompting, evaluating, and iterating while making real engineering calls about models, data, and architecture. The pipeline is the vehicle; the real subject is agentic engineering — designing systems where an AI agent does meaningful work, and making the judgment calls that separate something that demos well from something that actually runs.
Working knowledge of Python and APIs is all you need. No ML experience required.
How the course is structured
You'll move through the pipeline the way you'd build it in practice — ingestion, ideation, production — then layer in the refinements that make it robust: vector similarity scoring, image variety, error handling and retry logic. There's a bonus module on generating the audio overview, and a wrap-up that ties the architecture together. Lessons are kept short and focused, so you can work through a stage, build it yourself, and move on.
This is hands-on throughout. You're not watching someone narrate finished code — you're making the same decisions about models, data, and architecture that you'd face building this for real, and seeing the reasoning behind each one.
What makes this different
Most material in this space is either narrow how-to tutorials or high-level hype about what AI "will" do. This sits in between, where the actual work happens: a complete system, built end to end, with the engineering tradeoffs made explicit. You'll come away not just with a pipeline you built, but with a way of thinking about agentic engineering you can apply to your own projects.
You'll also get the complete reference implementation as a public repository, so you can compare your work against a known-good build at any stage, and access to the GammaVibe Community, where you can ask questions and work through your build alongside other students and me.
Build alongside me — real code, real decisions
Produce a finished newsletter, header image, and podcast
Understand the whole system, not just the parts
Curriculum
Who this is for
This course is built for senior engineers and independent builders who want to work with agentic systems at the level of architecture and engineering decisions, not surface-level tutorials. If you've shipped software, you'll feel at home.
You need working knowledge of Python and APIs. You do not need any machine learning background — there's no model training, no math prerequisites, no ML theory. The course is about engineering with AI, not building AI from the ground up.
It's also deliberately tool-agnostic. The build uses Claude Code, but the principles apply to whatever coding agent you prefer — this isn't a Claude Code tutorial, it's a course about how to engineer agentic systems, with a capable agent as one of your tools. The course walks you through setting up the services you'll build with — your coding agent, an AI model API (Gemini), and a news API — which charge modest usage fees as you go.
Want hands-on help too?
Get the course plus two 1:1 sessions with me to work through your own build.
Meet your instructor
Mirko Froehlich spent 13 years at Google, most recently as a Senior Engineering Manager on Google Drive. Before that, his career spanned more than a decade in the startup world — a startup veteran who has worked across many verticals, in many roles, and with many programming languages as a full-stack engineer.
Today he runs GammaVibe, his personal R&D lab and home of a daily newsletter that uses an autonomous AI pipeline to surface fresh startup ideas. He builds in public, sharing what he learns about agentic engineering on his YouTube channel and through GammaVibe Academy.
The Agentic Pipeline is his first course — a hands-on walkthrough of the same architecture that powers GammaVibe, built for engineers and independent builders who want to construct autonomous AI research pipelines of their own.
