Hi, I'm Austin.
21-year-old developer shipping production systems that most teams need 10 people for. I build AI-powered enterprise tools, then help companies do the same.
Updated February 2026
I did two years of a Law degree at the University of Birmingham. Took a gap year, looked at where AI was heading, and realised I'd rather build the future than argue about the past. Never went back. Best decision I've made.
At 19, I started building Change Radar — what became a 550,000-line enterprise SaaS platform. No funding, no team, no CS degree. Just me, TypeScript, and an unreasonable amount of caffeine. I presented it to the CIO of Allianz. I was the youngest person in the room by two decades.
Now I split my time between building my own products and helping professional services firms become AI-native. I don't give people chatbots and call it consulting — I teach teams to automate their own work, then get out of the way.
Ship fast, learn by building, stay honest about what you don't know.
Enterprise project intelligence — built solo, end to end
Multi-tenant PMO SaaS with 550K lines of TypeScript. 8 orchestrated AI engines handle risk analysis, automated reporting, and natural-language queries. 85+ database migrations, Row-Level Security, AES-256-GCM encryption, full CI/CD. Presented directly to the CIO of Allianz — youngest person in the room by 20 years.
Stop burning 25K tokens on schema lookups — distill once, grep forever
Claude Code plugin that distills your entire database schema into a grep-friendly index file. Instead of blowing 25K+ tokens on MCP calls every time the AI needs a table name, it searches a ~300-line local file. Zero tokens, instant results. Tested on Change Radar's 318-table schema — eliminated 32,000 tokens per lookup.
AI-powered document chasing for accountancy practices
Intelligent multi-channel escalation engine that chases clients for tax documents. Email → SMS → WhatsApp escalation with AI document recognition, magic-link client portal, and practice dashboard. Solving a £945M problem across UK accountancy — 500-750 hours per practice wasted chasing documents annually.
AI knowledge base for operations teams
RAG-powered internal knowledge base for manufacturing. Upload company SOPs, search with natural language. Workers ask questions, get answers from company docs instead of bothering supervisors.
Automated websites for tradespeople
End-to-end pipeline: scrape prospects from Google, generate bespoke websites, deploy automatically, cold-outreach with the finished site. Built the pipeline, ran the outreach, learned the sales lessons.
TypeScript · JavaScript · Python · Rust · SQL
React · Next.js 14/15 · Tailwind CSS · shadcn/ui
Node.js · tRPC · Drizzle ORM · Prisma · REST
OpenAI · Anthropic Claude · RAG · pgvector · Embeddings
PostgreSQL · Supabase · Neon
Vercel · Cloudflare · AWS R2/S3 · Docker · GitHub Actions
Clerk · Supabase Auth · RLS · AES-256-GCM
Stripe · Resend · Twilio · QStash · ElevenLabs
Coming soon
Coming soon
Coming soon
More writing coming soon. For now, find me on LinkedIn.
I help professional services firms — accountancies, consultancies, agencies — become AI-native. Not hype. Not chatbots. Real automation that saves real hours.
I'll map your workflows, identify where AI actually helps (not where it's hype), and give you a concrete implementation plan.
Need an internal tool, automation pipeline, or AI-powered feature? I'll build it. Fast. In production-grade TypeScript.
Hands-on workshops to make your team AI-native. Not slides and theory — real tools, real workflows, real results.
Building something? Hiring? Just want to chat? I'm always up for a conversation.