15 hours ago
Paris, FranceSenior
Responsibilities
- Build agentic pipelines using chains of LLM-based agents for complex clinical documentation tasks.
- Integrate LLM provider APIs into production features and develop AI capabilities that improve clinicians’ workflows.
- Design prompts and evaluation frameworks to measure and improve output quality.
- Write production-grade code and own features from design through deployment and monitoring.
- Collaborate with product managers and cross-functional squad members to deliver reliable, scalable product features.
Requirements
- At least 3 years of experience writing production code, with strong software design, testing, and system architecture skills.
- Proficiency in one or more backend languages, including Kotlin, Java, C++, or Python.
- Experience building features or products with LLM provider APIs, prompting, and evaluation; model training or fine-tuning is not required.
- Ability to turn ambiguous product requirements into working, testable pipelines.
- Strong communication skills and a practical, hands-on approach to software engineering and LLM development.
- Experience with multi-agent systems or agent orchestration frameworks is a bonus.
Benefits
- Stock ownership.
- 100% healthcare coverage.
- Meal vouchers.
- 50% public transportation cost coverage.
- Exercise classes during the workday, including yoga, running, pilates, and HIIT.
- Unlimited budget for book purchases.
- Paris-based hybrid work with 1–2 office days per week and the possibility of working remotely for 1–2 weeks once per year.
Tech Stack
Categories
About Nabla
Nabla builds an ambient AI assistant for clinicians that captures visits, generates notes, and streamlines related tasks through integrations with major EHRs. The software is sold to healthcare organizations and health systems, supporting workflows before, during, and after visits across specialties and languages. Founded in 2018 and headquartered in New York, Nabla is a privately held company focused on clinical documentation and point‑of‑care workflows.