
Greg Ceccarelli
477 Subscribers
Co-Founder • Ex-CPO Pluralsight • Data at GitHub, Dropbox & Google
Greg Ceccarelli will make you an agentic engineer
I co-founded SpecStory and have built our products (open source extensions and Stoa) entirely by steering agents with almost no hand-written code.
I wrote about it extensively: the Beyond Code-Centric whitepaper, Goal Engineering, and 25 Patterns in Agentic Engineering book.
Before this I was Chief Product Officer at Pluralsight and led data teams at GitHub, Dropbox, and Google. I'm passionate about teaching what it takes to actually ship with agents: the process, mindset and actual techniques that work.
Previously at



Courses
Multi-day, guided programs to get real results
Lightning Lessons
Free, interactive sessions to explore new topics
Alumni reviews
A foundational, well-paced and enjoyable course, highly recommend. The course material is essential for understanding how to bind harnesses around your agents such that they stay within the bounds of your long running tasks. Thank you, Greg!
Himanshu
Cohort 1
Data and Analytics Engineer · Altarum
Hardcore Agentic Engineering for Builders Who Ship lived up to its name in a very good way. Greg’s course changed and deepened the way I work with coding agents. The course is well structured, carefully sequenced, and clearly thought through; Greg takes a complicated subject, breaks it into manageable pieces, and helps students apply it to real work. He is an incredibly sharp and capable engineer who is eager to share what he has learned, while also being personable, friendly, and genuinely approachable. His investment in each student came through in every interaction, and I left with practical habits I’m already using and a much stronger sense of how to get real, trustworthy work from coding agents.
Jack
Cohort 1
CEO & Co-Founder · Comprendo
This is an excellent course. Greg is a knowledgeable engineer and an outstanding instructor. I came in wanting to build my own particular kind of harness, and he engaged with it directly, fitting it into the course frame and offering concrete suggestions to improve it.
The material is as current as this fast-moving topic allows. We did a deep dive on Prime Intellect's Prime Agent and on the DeepSeek Harness, which dropped on August 13 while the course was running.
One practical note: the teaching example is mostly TypeScript, and the parent harness is mostly Rust. If you want to build your harness in Python, expect a few extra hoops, though nothing difficult.
Recommended for anyone working on long-running agents that need to produce reliable results.
Randall
Cohort 1
CEO/CSO · Indigo BioAutomation





