Projects

Selected projects where I applied product thinking, automation, and QA principles end-to-end.

AI-assisted training plans for soccer coaches.

Problem: Coaches spend hours preparing sessions and struggle to adapt plans to team-specific goals, age groups, and training objectives.
Solution: An AI-powered web app that generates structured, ready-to-use training plans tailored to the coach's context, team, and objectives.
Role: Product owner, designer, and engineer. Scoped the experience, built the app, and defined the quality criteria.
Tech: Next.jsTypeScriptAI / LLMVercel
Result: Live in production; used to plan real training sessions.

Count anything, right from Telegram.

Problem: Recording recurring counts or events usually means opening a spreadsheet or an app — friction that breaks the habit.
Solution: A Telegram bot that captures counts and events through natural chat messages and stores them for later review.
Role: Solo build: product, backend, bot integration, and testing.
Tech: Telegram Bot APINode.jsCloud-hosted
Result: Live and actively used for personal and side-project tracking.

Telegram-style todo list on the web.

Problem: Most todo apps feel heavy and disconnected from the fast, chat-like flow people already use every day in messaging apps.
Solution: A web-based todo list that mirrors the Telegram experience — quick capture, lists, reactions, edits, replies, and search — so managing tasks feels like chatting.
Role: Solo build: product, design, frontend, and backend.
Tech: Next.jsTypeScriptVercel
Result: Live at todolist.delrayo.tech for daily personal task tracking.