AI Product Engineer, Clinical Tools
Knownwell›
📍Remote
Posted Today · via lever
Apply on lever→Job Description
👋 Meet knownwell, weight-inclusive healthcare for all. Join a dynamic company that is changing the way care is delivered to patients with obesity. knownwell is a weight-inclusive healthcare company offering metabolic health services, primary care, nutrition counseling and behavioral health services for anyone of any size. Our hybrid model allows for both in-clinic and virtual care to bring support to patients where and when they need it.
We’re seeking a rare kind of builder: an AI Product Engineer, Clinical Tools who can move fluidly from defining what to build to actually building it. This role sits at the intersection of applied AI engineering and clinical product development, with ownership of both the product vision and the technical execution for our clinician-facing AI tools. The initial flagship is our Clinical Decision Support product, powered by a RAG pipeline over clinical knowledge sources.
This is not a role for someone who wants to hand off a spec and wait — or receive one. There is no PM layer above you. You’ll define what gets built, build it, and own the outcomes. Day-to-day that means being hands-on in the codebase — designing and iterating on RAG pipelines, evaluating model outputs, tuning retrieval strategies — while also setting the product roadmap, partnering directly with clinicians, and aligning stakeholders across the organization. You’ll report directly to the Chief Product Officer and are expected to operate with a high degree of autonomy.
Responsibilities:
-
End-to-End Product Ownership — Define and own the product vision, strategy, and roadmap for Clinical AI tools with no PM layer above you. Translate clinical workflow needs into prioritized, sequenced plans and own those decisions through to shipped features — balancing near-term delivery with longer-term strategic bets.
-
RAG Pipeline Design & Iteration — Architect, implement, and continuously improve the RAG infrastructure powering clinical decision support: chunking strategies, embedding models, vector database design, retrieval and reranking approaches, and evaluation frameworks.
-
Prompt Engineering & Model Behavior — Design and iterate on prompt strategies, system instructions, and guardrails to produce reliable, clinically appropriate outputs. Build evals to measure quality systematically, not just anecdotally.
-
AWS Infrastructure — Deploy and maintain scalable, HIPAA-compliant AI infrastructure on AWS. Make informed tradeoffs between managed services and self-hosted components, weighing cost, latency, compliance, and performance.
-
Clinical Data & Integration — Work with healthcare data sources (EHR-adjacent, structured and unstructured clinical content) and integrate AI outputs into clinical workflows in ways that are accurate, auditable, and safe.
-
Clinical Stakeholder Partnership — Own the relationship with clinicians and clinical operations directly — there is no PM intermediary. Build deep understanding of clinical workflows, pain points, and the real-world constraints of care delivery, and translate that into product and engineering decisions.
-
Technical Collaboration — Partner with backend and frontend engineers to integrate AI capabilities into the broader product. Contribute meaningfully to architecture discussions and help the team make sound infrastructure decisions.
-
Staying Current — Actively track developments in applied AI research, RAG techniques, and agentic workflow design. Bring relevant advances to bear on the product — whether that's a new retrieval method, an emerging evaluation framework, or a better pattern for building reliable multi-step AI pipelines. We expect you to know what's happening in the field and have a point of view on what matters.
-
Stakeholder Communication & Alignment — Translate technical AI concepts and tradeoffs for clinical and operational stakeholders. Align engineering, clinical ops, and leadership around priorities and progress without requiring others to fill in context. Write requirements and documentation that others can act on independently.
Requirements:
-
5–8 years of experience in software or AI/ML engineering, with a meaningful portion in applied AI product development
-
Hands-on experience building and operating RAG systems in production — you’ve made real decisions about chunking, embeddings, retrieval design, reranking, and evals, not just prototyped them
-
Strong Python skills; comfortable building pipelines, writing evaluation harnesses, and iterating on model behavior programmatically
-
SQL proficiency sufficient to query data independently, pull your own product metrics, and answer analytical questions without waiting on a data team
-
Experience deploying AI workloads on AWS; familiarity with relevant services (e.g., Bedrock, SageMaker, Lambda, RDS/Aurora, S3) and the tradeoffs between them
-
A product mindset — you think about user problems and outcomes, not just technical implementation, and you can write a clear spec as readily as a pull request
-
Experience with API design and integration, and the ability to collaborate closely with frontend and backend engineers without being a bottleneck
-
Healthcare or regulated-domain experience preferred; you understand why accuracy, auditability, and safe failure modes matter more in clinical contexts than in most
-
Familiarity with LLM safety tooling (e.g., guardrails, output validation frameworks) and an instinct for where AI systems can fail quietly
Details
- Department
- R&D
- Work Type
- remote
- Posted
- April 13, 2026
- Source
- lever