Most AI health companies are still pitching what they'll build. Beni is already live inside primary care practices, handling real patient interactions through cutting-edge voice AI and multi-modal messaging, coordinating care between visits, and generating meaningful clinical outcomes. What we're building next is the agentic layer that makes all of that dramatically more capable. We're hiring an engineering lead to build it.
About the Role
This is a hands-on engineering role at a company transitioning from a strong AI-assisted product to a fully agentic one. You'll architect and build the systems that power Beni's next generation: multi-modal patient communication (voice, SMS, structured data), intelligent care coordination workflows, and the integrations that connect our platform to the clinical data that makes it useful.
You'll work across a founding team that includes highly regarded medical experts with deep expertise in the space, alongside an established, highly accomplished product, sales and operations executive team. You're the person who makes the vision real.
What You'll Own
- Architecture and implementation of Beni's agentic AI systems: orchestration, memory, tool use, multi-step reasoning
- The multi-modal communication layer, including our voice AI pipeline. Real patients are having real clinical conversations with our system today. You'll make that better, faster, and smarter.
- EMR integration work: connecting Beni's intelligence to the clinical data that lives in practice systems (FHIR, HL7, and whatever else the practice world throws at you)
- Production reliability for AI outputs in a clinical context: evaluation frameworks, monitoring, and graceful failure handling
- Backend infrastructure that scales with our practice base without becoming a liability
- Close collaboration with product to ship things that work, not just things that demo
Who We're Looking For
- Deep hands-on experience building AI systems in production: LLMs, RAG, agentic orchestration, voice AI, or some combination. You've shipped real things to real users.
- Strong backend fundamentals: APIs, async systems, data pipelines, infrastructure that doesn't fall over
- Real experience with or genuine appetite for healthcare data standards and EMR integrations
- Judgment to know when to build and when to integrate. We're not a research lab.
- You stay current with AI development tools and actually use them in your daily work. You're fast because you've invested in how you work, not because you skip steps.
- Exceptional written communication. This is a distributed, async team and clarity is load-bearing.
- Self-direction: you surface blockers early, move without being pushed, and know when to ask
The Opportunity
We're putting agentic AI to a real test. We have live patients, live practices, and a tight feedback loop between what we build and what actually happens in care. You'll be rebuilding a working clinical interaction system into the agent layer that supports continuous care between visits. The constraints are real, the stakes are real, and the impact is measurable.
This is the kind of engineering problem that doesn't come along often. Real voice AI in clinical settings, real agentic workflows with real consequences, and a product that's already operating, not waiting for product-market fit.
This role is open globally. We hire for skill and judgment, not geography. Compensation includes early-stage equity. We're a small team building infrastructure for a problem that matters.