AI Slop Prevention Workshop

Description

Here's the message for coding with AI: in 2026, the problem is mostly you. And that's okay, we're all still very early in this game.
This workshop event is for anyone simultaneously enthusiastic about the promise of agentic engineering and fatigued by the reality of the AI slop slog.
In this two-hour workshop, we're going to demonstrate how a few of the best practices we've learned so far translate into obviously better code. Cutting to the chase:
Scope your sessions / PRs sanely and stick to it
Enforce opinionated linting and type checks to establish a baseline level of code quality
Write one foundational end-to-end integration test to ensure no effort totally derails functionality
Always open PRs against an upstream branch and require CI green checks before merging
Recognize software fundamentals still drive code quality, set standards and be self-aware about your skillset
After demonstrating the value of utilizing these in an agentic engineering workflow, every attendee will apply them with their own devices in their own codebases.
Prerequisites
Experience / comfort with Claude Code CLI
Some web development experience
Some TypeScript experience
Some GitHub experience
What you'll walk out with
By the time you leave, you'll have:
At least one of the five patterns actually enforced in your own repo — not "I'll do it Monday," but committed and pushed
A PR open against your own main with green CI, if you sprint
A one-page cheat sheet with the exact configs and prompts that make these patterns work
The vocabulary and the receipts to call slop slop in your next standup
Who this is NOT for:
People who haven't used Claude Code yet. Try it for a week first — this won't be your first stop, and you won't get $ value out of it
People looking for a "should we use AI" debate. We've moved past it. So have you, probably
Anyone hoping for a watch-and-listen session. Bring a laptop and get ready to jam or you'll be staring at our screens for two hours
Anyone whose never looked at TypeScript code. We're going to be specific about the stack and pretending otherwise would be slop
Presenter: Geoffrey Sechter
Geoffrey runs Lightstrike Labs (this workshop's sponsor), an AI-first consultancy that mostly gets called in after Claude Code has had its way with someone's repo. TypeScript, React Native, general slop archaeology. Has personally shipped, deleted, and re-shipped more agent-generated code than is probably healthy.

Sponsors

Lightstrike Labs: AI-first product engineering consultancy