The long version.
A short tour of the work, the route here, and what I’m chasing now.
Came up the Swiss way: a 4-year apprenticeship at UBS. Year one was , covering IT fundamentals, Scrum, and group projects. The next three years I spent in Wealth Management on trading applications, building Java REST services and internal trading interfaces. That stretch taught me how high-stakes systems actually get built and shipped.
Since 2022 I’ve been on the Conversational Banking team, working on the , UBS’s E-Banking and Mobile Banking chatbot. I started on the Java backend, owning services that powered the bot in production, and drove the Java side end-to-end into its first production release in 2023. I was on the team before generative AI hit; when it did, I helped drive LLM integrations and adoption, shipping one of UBS’s first external GenAI use cases. In 2024 I moved into fullstack by building the , a React and Python platform the team uses to monitor, analyze, and test the bot end-to-end, and shipped the first test automation suite in the Conversational Banking crew along the way. An internal rotation into that year gave me a first taste of NLP and shipped a new REST service into production. That same year I also won the with a cross-team squad, and was promoted to the following spring.
Right now I’m most interested in agents and applied AI, tooling that makes developer workflows faster and day-to-day work less tedious. The Chatbot Portal is one example, and I want to do more of that. Currently shipping evaluation and data-labelling tooling for the Virtual Assistant, tinkering on side projects in the evenings, and going deep on agent architectures.
— Y.
§ Stack
- Daily
- Python· TypeScript· React· FastAPI
- Backend & infra
- Azure· OpenAI· Postgres· Redis· GitLab CI
- Exploring
- LangGraph· Agent tooling· Cloudflare Workers