Role overview
We are Lukas, Stefan & Sean, the co-founders of Mercura. We share a deep commitment to excellence and the discipline to pursue what others might consider impossible. Two of us come from non-academic families, and we all know what it means to create our own paths and overcome barriers. This experience shapes our belief in equal opportunity and fuels our dedication to social mobility.
At Mercura, we believe that the best ideas emerge when people from different backgrounds collaborate together. We hire based on ability, mindset, and values, and we live by one simple rule: the best idea wins.
We actively support initiatives that promote diversity and gender equity. Each year, we also commit a portion of our personal income to organizations working to expand opportunity for others.
We’re a small, fully in-person team based in Munich that feels like family. We bring extraordinary people together and deeply invest in each other’s growth as mentors, teammates and friends.
We care, trust, and support one another, and we make sure that joy and collaboration is at the heart of everything we do.
What you'll work on
- Work across the full stack to build AI agents end-to-end
- Work with embeddings and fine-tune LLMs for classification and reranking
- Optimize algorithms for product search
- Apply RLHF & DPO to align LLMs with human feedback
- Build data pipelines to process large-scale unstructured data efficiently
- Make our backend scalable, stable and efficient
What we're looking for
- High agency and experience with building software and iterating quickly
- Experience with NLP, LLMs, Embeddings, ML, and AI-powered applications
- First Experience with full-stack development (TypeScript, Python, React, Next)
- Strong problem-solving capabilities and willingness to learn new technologies
- You’re excited to strive for excellence and build software in a fast-paced startup environment
- You want to work in-person in Munich with more intensity than in a 9-5 job
Interview process
A take-home technical challenge to showcase your skills, followed by a call to walk us through your approach, insights, and solution for the take-home exercise.