Prompt Engineering & Pipeline Management: Design, iterate, and maintain prompt systems that reliably generate high-quality clinical documentation across a growing number of templates.
Build scalable prompt architectures that let us add new output formats without fragmentation.
Evaluate & Integrate AI Models: Benchmark and select the best LLM solutions for clinical documentation. Stay on top of the rapidly evolving model landscape and assess new options.
Build Evaluation Pipelines: Design practical evaluation processes across our data pipeline from audio and transcription to LLM-generated output. The goal isn't perfection but knowing when output is good enough.
Ensure Clinical Output Quality: Work closely with our psychologists and product team to define what "good" looks like, ensuring features meet the real-world expectations of users.
Experiment, Ship, Iterate: Design experiments and data labeling strategies, push them to production, measure real-world performance, and use the insights to drive the next iteration.
Requirements
Prompt Engineering & LLM Expertise: 2+ years of deep experience in systematic prompt engineering, building structured prompt systems that produce reliable outputs at scale.
Software Engineering Foundation: Solid engineering skills in Python and cloud infrastructure (we run on Azure). You build reliable, production-grade systems, not just notebooks.
German Language Skills: You need to understand Swiss German to sanity check AI-generated notes against the original audio. Strong High German for evaluating written clinical output.
Pragmatic Startup Mindset: You know that good enough is good enough. You ship fast, iterate based on real-world feedback, and resist over-engineering. Ideally you've worked in an early-stage SaaS company.
Clear Communicator: Our team works in English day-to-day. You can articulate technical decisions and collaborate effectively across functions.
Tech Stack
Azure
Cloud
Python
Benefits
Real Impact, Fast: 500+ therapists use our product daily, your work improves their lives within days, not quarters.
Unique Problem Domain: Clinical AI at the intersection of Swiss German speech, mental health, and regulatory requirements.
Massive Ownership: A team of ~10 means your decisions shape the product directly.
Flexibility: Remote-first with flexible hours and a coworking space in Zürich when you want it.
Equity: Meaningful stake in a high-growth healthtech company.
Senior AI Engineer, Healthcare Startup at PlaynVoice | JobVerse