Learn everything about context engineering
An eight-week LLM course interface built around Conway's Game of Life: the lesson map is the simulation.
Awwwards Honorable Mention.
Eight weeks of lessons is a lot of navigation to make legible. Building the map as a Game of Life board meant you could see the whole course at once, and people poked at it instead of skimming past.
Related posts
Interim Engineering Lead — RaceMake.com
I led architecture and delivery for an LLM-powered sim-racing telemetry platform: a 
Rust/
Tauri desktop app and a 
React web app.
- I rebuilt both clients around extensible high-frequency telemetry pipelines and per-game adapters.
- I built an AI pit team and interactive 3D visualisations on top of the live data.
- I was brought in to lead, not to take direction. I set the architecture, then shipped it with the team.
TypeScript
Next.jsLead Product Engineer / Consultant — 
WithSeismic
I've run an independent product-engineering consultancy for 10 years and delivered 100+ engagements, for clients including 
Framer, 
The Motley Fool, 
Groupon, 
Contra and Weedmaps.
- Consulting, scoping, pricing, delivery and ongoing support for technical builds.
- Product engineering, growth engineering, technical marketing, analytics plumbing and interim engineering leadership.
…
Most engagements start on someone else's codebase, with a date already promised to a customer. Ten years of that has taught me more about what makes a product hold up than any of my own builds did.
Founding engineer — 
GrowthMatch
Video-first content operations for B2B, at a 
Techstars alumni company: a couple of hours of expert input a month comes out as a full campaign of video, ads and written assets.
- I built the product as founding engineer — the generation pipelines and the application around them.
- LLM plumbing doing the work an editor would otherwise scrub through by hand, wired into the campaign and distribution surfaces.
The intro video at the top of this page was cut with it. That is the product test I actually trust — I reach for it because it is faster than the alternative, not because I wrote it.
LLM-assisted video editing, for Snacker.ai
A pipeline that takes a raw recording to a polished, cut video in under 20 seconds — transcription, edit decisions and rendering, with the model doing the part an editor would normally scrub through by hand.
The interesting constraint was latency, not quality. Anything over about half a minute and people go and do something else, so the whole thing had to be fast enough to feel like a button rather than a job you submit.
MIT Generative AI and LLM course
Product Hunt — #1 Product of the Week
Performance Marketing Awards