Full-stack developer with Java / Spring Boot on the back and React / TypeScript on the front. I combine technical skills with business understanding, hands-on AI development from an agency and 1,100+ workshops delivered.
Spotted the need on-site. Built the MVP solo. Client took it forward.
An AI voice assistant for electricians. Call the bot, talk about the job, reports fill themselves. Electrician approves. From shadowing an electrician through a full work day to MVP handover to the senior dev team after client paid to take it further.
200+ registered matches in two weeks. Simple UX led to heavy usage.
A full office ping-pong league with AI-generated character portraits for every player. TypeScript frontend, Spring Boot backend, PostgreSQL. Deployed and organically adopted.
Real integration. Real companies. Real receipts.
Photo a receipt, review the AI-suggested booking, approve. Integrated directly with Bokio's API. Used by real companies today.
Made working out easier. With agentic workflows.
Call an AI trainer and get a real workout on the spot. Receive instructions, coaching and ask questions. Agentic flows with tool use under the hood. I'm product owner and UI/UX lead in a team of five.
Spotted the need on-site. Built the MVP solo. Client took it forward.
Electricians are often behind on filling in reports on jobs they perform. Late reports are more prone to errors and cause problems for invoicing. It's also a task electricians don't enjoy doing. I identified this problem at a meeting with the client on day one at GNR8. I then spent a full day in the field shadowing an electrician in order to build an AI voice assistant: call the bot, talk about the job, and it fills in time and activity reports automatically. The client paid for the MVP and for taking it forward and I handled handover to our senior dev team for integrations with their internal systems.
200+ registered matches in two weeks. Simple UX led to heavy usage.
A full office ping-pong league for SALT. Every player gets an AI-generated character portrait. Matches, leaderboard, and stats. All live on Railway. The real measure: colleagues started using it on their own and kept coming back.
Real integration. Real companies. Real receipts.
Tekont connects directly to Bokio's accounting API. You photograph a receipt, the app reads it and suggests a booking, and you approve it. No manual data entry. Currently used by a small number of companies in the arts sector and actively being adopted by others.
Made working out easier. With agentic workflows.
An AI trainer you actually call. Simpler than dealing with menus with endless lists of exercises and more fun than doing the usual routine. Choose your trainer out of a cast of characters with distinct personalities.