AI Engineering
End-to-end AI solution delivery: I gather requirements with the business, design the architecture, build the UI and backend, and lead rollout and change management. Technical work spans RAG, agentic systems, orchestration, retrieval and evaluation, using local open-source models or cloud APIs to fit the data and infrastructure.
Automate routine coordination with agentic systems while keeping people in control of consequential decisions.
Make internal knowledge searchable and permission-aware, with retrieval that respects who is allowed to see what.
Run confidential workloads on local open-source models instead of sending data to a public API.
Define requirements and architecture
I work directly with business stakeholders to understand the workflow, agree acceptance criteria and map data access and human approvals. A PoC or MVP tests the architecture before a larger investment.
Build and iterate with users
I implement the model workflows, APIs and interface, combining LLM orchestration with deterministic code. User feedback drives changes to both the UI and backend; evaluation checks whether the system is useful and reliable.
Lead rollout and adoption
I take responsibility for deployment, user training and change management. Clear failure paths, technical documentation and continued iteration support the transition from working prototype to everyday use.
Do you only build the technical solution?
No. I own the path from business requirements and solution architecture through hands-on UI and backend development to rollout, training and change management. I work directly with users and decision-makers throughout delivery.
What have you actually built with LLMs?
A secure RAG platform for a private bank where each department runs its own governed GPT behind one shared chat interface, with authorization checked before retrieval. An agentic recruitment MVP whose tested happy path fell from roughly five hours to about five minutes, using orchestration and sub-agents with humans on key steps. This was MVP testing, not a production-wide result. A system that turns a plain-language process description into a valid BPMN 2.0 model. And a Swiss German transcription tool with LLM summaries tailored per department.
Do you work with local models or cloud APIs?
Both, with the focus on local. Most of my work uses open-source models on self-managed infrastructure, which is what makes confidential documents and internal processes workable at all in regulated environments. Cloud models are used where the data allows it and the capability is worth it.
Do agentic systems remove the human entirely?
No, and they should not. In the recruitment system almost everything around the important steps is automated, but a person still approves the decisions that carry consequences. Autonomy is applied where it is safe and withheld where it is not.