About this role
Summary: Build and improve LLM agent workflows for nursing documentation and care planning. Own model evaluation, investigate failures, and train/adapt models to ensure accuracy, speed, and privacy for care professionals.
🎤 Why voize?
At voize, we believe the greatest gift to frontline workers is time: time to care, connect and be present. Our AI helps care professionals turn spoken notes into structured documentation, so they can spend less time on administrative work and more time with people.
More than 200,000 care workers use voize. Join us to build technology that makes a difference in their daily work.
💡 Your mission
Build agents that nurses can trust: accurate, fast and safe enough for real-world care. You’ll combine hands-on ML with thoughtful engineering and rigorous evaluation, using production data and feedback to improve the product.
We’re hiring four ML engineers through one process. Depending on your strengths and interests, you’ll work on hospital documentation, US senior care, German care agents or the shared NLP platform. You’ll be part of the NLP chapter and work closely with product and engineering colleagues.
🚀 What you’ll do
- Build and improve LLM agent workflows for nursing forms, assessments, care reports and care-plan suggestions, working on tools, context, state and prompts.
- Own evaluation for your area: create benchmarks from production traces and synthetic scenarios, build regression tests and use results to inform what ships.
- Investigate real-world failures, reproduce issues across the app, backend and models, and turn what you learn into fixes, tests and data-labeling tasks.
- Train or adapt models where they outperform rented frontier models on quality, speed, cost or privacy—for example, for personal-data detection or smaller, faster models.
- Make latency, cost and privacy part of your engineering decisions, from model routing and caching to protecting sensitive data.
- Partner with product, engineering, data and customer-facing teams, and use feedback from care professionals to guide improvements.
🤝 What you bring
- Experience training, evaluating and shipping ML or NLP models to production, and explaining the decisions and results behind your work.
- Experience building evaluations for ML or LLM systems and using error analysis and regression tests to guide product decisions.
- Hands-on experience with LLMs in production, including agents or tool use, structured outputs and context design.
- Strong software engineering skills: you write tested, reviewable production code and can learn an unfamiliar language or framework.
- An end-to-end approach to messy problems, sound judgment with sensitive data, and the ownership to work through ambiguity in a fast-paced product team.
- Fluent English.
🛠️ Our stack
Our ML and evaluation tooling uses Python; our agent and app code uses Kotlin Multiplatform and JetBrains Koog. We use Claude through our inference service, alongside open and on-device models where they make sense. You don’t need to know our specific languages or frameworks before joining: depth and the ability to learn matter more than a keyword match.
🌱 What you can expect
- Work on a product used by care professionals, with direct feedback and a clear social impact.
- Shape a young agent stack and grow into ownership of a domain, market or technical area.
- Do real ML work, from training and evaluation to production, alongside collaborative engineering and product teams.
- Join a team that values written communication, shared learning and regular time together.
- Become a co-creator of our success with stock options and access to language courses.
✨ Ready to talk?
Apply to join us in giving care professionals more time for people. We look forward to hearing about the systems you’ve built, what you learned and the impact they had.