mnemic · Posted Nov 6, 2026
Tech Lead / Founding AI Engineer
Zurich · Part-time, 16 h/week · Revenue-based pay + equity
We're mnemic, two founders from ETH Zurich. Our product Novalis answers compliance questionnaires – due diligence, vendor risk – and links every answer to the exact passage in the client's own documents. It works in German and English and runs on-premise or in the client's Azure tenant, with a locally hosted language model.
Pilots run with one of Switzerland's largest law firms and soon with a Swiss asset manager. We're looking for a Tech Lead to own architecture, deployment and AI quality with us.
Possible projects
Prove quality with numbers. Evaluate Novalis on established legal benchmarks (e.g. LegalBench-RAG) and our own German/English test sets. Fine-tune embedding, reranking or language models where it measurably helps.
Ship on-premise as a product. Turn our deployment into versioned, installable releases: Docker images, configuration, updates and monitoring that work inside a client's own network.
Make Novalis a standard Azure offering. Deployable into a client's Azure tenant in hours, not weeks – infrastructure as code, Entra ID sign-in, private networking, and possibly an Azure Marketplace listing.
Turn client requirements into features. Work directly with legal and compliance teams and turn their feedback into reliable, user-friendly features.
Own the architecture. Make the key technical calls with the founders: security, testing, CI/CD and what we build next.
What you bring
Must-have
- You've built a RAG or LLM application that real users or a real evaluation put to the test – at work, in a university project or on your own.
- Solid Python and Docker, and you've deployed and maintained something in production, including logs, updates and backups.
- You take ownership and are comfortable making decisions with incomplete information.
Nice to have
- Microsoft Azure – or AWS/GCP and the drive to go deep on Azure.
- Serving open-weight LLMs on GPUs (e.g. vLLM) and RAG evaluation tools (e.g. Ragas, DeepEval).
- Fine-tuning embedding or reranking models.
- Frontend or UI/UX experience.
- German – most of our clients' documents are in German.
- Interest in security and compliance standards (e.g. ISO 27001, FINMA requirements).
Students with strong personal or university projects are explicitly welcome.
What we offer
- Frontier AI in top-tier firms. You work first-hand with leading Swiss law firms and financial institutions and deploy frontier AI into their environments – not a demo, but production use on sensitive data.
- A founding role. Real ownership of architecture and technical direction, with a direct say in product priorities.
- A desk in Oerlikon. If wanted, on the 11th floor of Andreas Tower.
- Equity. Terms are finalised when we incorporate as a GmbH, planned for early 2027.
Compensation
We are pre-revenue and can't pay a fixed salary yet. Until we raise funding, 100 % of revenue after direct costs goes to compensating you and one co-founder, in a split and payment schedule agreed in writing before you start. Early revenue will be modest, so this role suits someone who is in it for the ownership and the experience.
How to apply
Send a few sentences about what you've built and why this role interests you, along with your CV and links to your GitHub, portfolio or relevant projects, to hiring@mnemic.ch.
- 12–24 hours per week, starting in December
- A short call, then a technical conversation about one of your projects