About the role
This is a hands-on role combining research with implementation. You’ll investigate model behavior, build reproducible experiments, and turn promising approaches into reliable software. The work includes language understanding, information extraction, longitudinal context, and evaluation, with opportunities to explore fine-tuning or training where there is a clear need and an approved data foundation.
VitalSyncLabs is building a health-context platform to help people understand what happens between medical visits. Our work connects personal experiences, health measurements, and history over time while preserving where information came from and what remains uncertain.
We’re an early-stage company based in Madison, Wisconsin. We bring together software, machine learning, and medical knowledge to build tools that support people and clinicians in understanding health—not replace their judgment.
Apply with your résumé. You may also include a GitHub profile, portfolio, LinkedIn profile, publications, or other relevant work. Please share only material you have permission to disclose; do not include patient information or confidential employer research.
What you’ll do
- Implement and evaluate model workflows that interpret information, connect relevant context, and preserve uncertainty.
- Compare models, prompts, retrieval methods, and processing approaches against clearly defined evaluation cases.
- Investigate failures involving missing information, conflicting evidence, inconsistent outputs, or unsupported conclusions.
- Explore fine-tuning or training approaches using data approved for that specific purpose.
- Work with software and medical-evaluation contributors to improve quality, runtime performance, and release readiness.
- Reproducible experimentation and evaluation workflows, including dataset versioning, regression tests, and experiment tracking.
What we’re looking for
- Practical experience developing or evaluating machine-learning systems through research, industry work, or substantial independent projects.
- Strong Python skills and familiarity with a modern machine-learning framework.
- An understanding of experimental design, model evaluation, data quality, and leakage prevention.
- The ability to turn an open-ended question into a testable hypothesis and explain what the results do—and do not—show.
- Careful software-engineering habits, including testing, documentation, and reproducible work.
Nice to have
- Experience with language models, time-series data, model serving, or fine-tuning would be helpful. Health-technology experience is welcome but not essential.
How we hire
We review applications for relevant experience, skills, and interest in what we’re building. If we move forward, we’ll arrange a conversation about your background, the role, and working at VitalSyncLabs. Depending on the position, we may discuss previous projects, research, or how you approach relevant problems. You’ll have time to ask questions, and we’ll explain any additional steps in advance.