Eric Chamberlain spent 25 years in education. Then he lost his job. Instead of sending CVs, he opened a laptop and started building.

Four months later, he'd shipped five apps — including Save Veritas, an oral assessment platform that tackles one of the biggest problems in schools right now: how do you know a student actually did the work when AI can do it for them?

In this episode, Eric walks us through everything — the origin story that started with his wife needing a French speaking app in Kuwait during COVID, the interview with an AI system that sparked the idea, and the technical journey from Bolt DIY frustrations to a robust, multi-assessment platform built on Vercel, Supabase, and the latest AI models.

We get into the real stuff: the tech stack choices, why PRDs and user stories changed everything, how to handle security and GDPR when you're not a developer, the tools that actually work on a bootstrap budget, and why boring, disciplined building beats vibe coding every time.

Whether you're a teacher curious about building your own thing, a solo founder figuring out your stack, or just fascinated by how fast non-coders can ship real products with AI right now — this one's worth your time.

In this episode we discuss:

• The assessment integrity crisis and why oral assessment solves it
• Building on free tiers — Vercel, Supabase, OpenAI, and Google Gemini
• Going from idea to PRD to user stories to shipped features
• Git branching, work trees, and protecting your codebase from AI agents
• Red teaming your own product for security
• Anti-Gravity, Claude Code, and Olama Cloud for budget AI-assisted development
• Aligning with UK DfE AI product safety standards
• The difference between vibe coding and actually building something that works

Connect with us:Alex Gray: https://www.linkedin.com/in/alexandergray84/

Darren: https://www.linkedin.com/in/darren-coxon/

Eric Chamberlain: https://www.linkedin.com/in/ericchamberlintech/