Product Manager with 5+ years across fintech, e-commerce, and startups — most recently at Banxware, where I moved from venture building into full product ownership. In 2026, I went deep on the technical side: AI software development, AI product building, and data analytics at Le Wagon. Below is what I built. Background in Psychology (B.Sc.) and an MBA in Finance & Marketing.
1. Daily AI Journal Bot
n8n · Telegram · OpenAI (GPT-4, Whisper, image generation) · Notion
A Telegram-based journaling companion that turns daily reflection into a lasting habit. Every morning it runs a guided five-prompt check-in — mood, gratitude, dreams, micro-victories, intention — accepting text or voice notes (transcribed via Whisper), then generates a structured summary, a personalized affirmation, and a unique AI image, all saved automatically to Notion.
The engineering challenge: n8n can't natively pause for a reply, so I built a stateful conversation loop with resume URLs stored in a lookup table, enabling true multi-turn dialogue. Design rooted in behavioral psychology — minimal friction, a well-timed cue, a familiar interface.
My first n8n project; built in a couple of evenings and ran every morning until my n8n trial expired — self-hosting it is next on the list.
https://www.youtube.com/watch?v=ZIpYAGTFZqM
2. Course Companion — AI Product Builder
Softr · Airtable · Supabase · n8n · OpenAI API
A learning platform with an AI assistant that actually knows the course. Built as my AI Product Builder capstone: Softr frontend, Airtable database, and a RAG pipeline — course content chunked, embedded as vectors in Supabase, and retrieved at query time — so the chatbot answers grounded in real course material, not generic knowledge. It also has a personalised conversation history (each chat securely attributed to the logged-in user via metadata) and a community forum where every new post triggers an automated, context-aware AI reply. n8n orchestrates all of it. I owned the full arc: problem framing, KPI definition, POC architecture, build, and documentation.
→ Full technical walkthrough (7 min)
3. Berlin Rental Market Dashboard
BigQuery · SQL · dbt · Data Studio
An interactive dashboard analyzing Berlin's rental market — my Data Analytics bootcamp capstone. Real-world listing data extracted and transformed with SQL in BigQuery, modeled into clean data layers, and visualized in Data Studio for exploration by district, price, and trend.
4. Beyond Exchange
Ruby on Rails · PostgreSQL · OpenAI API
A web platform for German teenagers (14-16) preparing for student exchange — built in a team of four during Le Wagon's AI Software Development bootcamp, from pitch to demo day in two weeks. The problem: exchange journeys suffer from fragmented communication, no peer community, and personal development left to chance. The Beyond Exchange App structures the journey, builds community, and centralizes communication for students, parents, and host families — the first digital pedagogical framework that makes high-touch exchange support measurable and guided.
→ GitHub