Simbie Health is on a mission to solve discharge. Transitions of care - any time a patient moves from one medical facility to another or to home - is fraught with vulnerability and many patients fall through the cracks. By developing the a tech-enabled platform focused on intercepting patients at discharge, we aim to ensure every patient transitions smoothly from their hospitalization to their homes. Working at Simbie, you’ll have the chance to innovate at one of the single most fraught and costly moments for patients in US healthcare: the transitions.

Simbie is founded by Natalia Khosla MD MSc, a physician from UChicago and Fulbright Scholar in Gender & Healthy Equity; Rachel O'Driscoll, a healthcare operations leader who ran US operations at Carrot Fertility and launched a tech-enabled clinic at Cayaba Care; and Natalia Perina, a software engineer from Princeton who led the patient recruitment product and data teams at TrialSpark.

About the Role

We’re hiring an AI Prompt & Product Engineer to help turn cutting-edge AI into real impact in healthcare. You’ll be responsible for shaping how our platform thinks, speaks, and solves problems — from designing prompts and fine-tuning models to scoping MVPs and testing new AI workflows.

In an early-stage startup, no two days look the same. Some days you’ll be experimenting with multi-agent setups or refining intake flows; others you’ll be pressure-testing our platform, breaking it on purpose, and helping translate ambitious client ideas into something we can ship in weeks, not months. If you enjoy blending engineering, product, and experimentation — this role is for you.

What You’ll Do

Prompt Design & Model Development

• Craft and refine prompts that support healthcare tasks like scheduling, eligibility checks, patient education, and intake.
• Contribute to fine-tuning by preparing datasets, running training jobs, and evaluating results.
• Explore new model behaviors and compare approaches across LLMs.
• Build internal tools and scripts to test and benchmark prompts at scale.

Product Discovery & Workflow Design

• Translate broad client requests into focused, testable solutions.
• Work with clinicians and staff to uncover bottlenecks and identify the core workflows worth automating.
• Break down complex problems into phased rollouts that let us learn quickly while reducing risk.
• Prototype new use cases and validate them with real users.

Testing & Quality Assurance

• Be the first user and toughest critic of our AI platform.
• Design testing scenarios that catch edge cases and failure modes early.
• Define metrics to measure conversation quality, accuracy, and user trust.
• Stress-test new features, models, and integrations before release.

Engineering & Technical Leadership

• Contribute directly to product code with Python/JavaScript for automation, integrations, and evaluation pipelines.
• Partner with engineers on architecture and deployment decisions.
• Identify opportunities to apply emerging AI techniques (RAG, multi-agent orchestration, lightweight fine-tuning).
• Explain technical trade-offs clearly to both internal and external stakeholders.

What You Bring

Core Experience

• 2+ years in AI/ML engineering, product engineering, or similar roles (startup or fast-paced environments a plus).
• Hands-on experience with LLMs, prompt engineering, and evaluation.
• Proficiency in Python and/or JavaScript/TypeScript.
• A track record of turning complex requirements into shipped products.
• Comfort with hypothesis-driven development and rapid iteration.

Technical Background

• Knowledge of AI frameworks, APIs, and fine-tuning techniques.
• Understanding of conversational AI design and multi-agent systems.
• Familiarity with A/B testing and metrics-driven decision-making.
• Ability to collaborate closely with engineers on system design and implementation.

Nice to Have

• Background in healthcare technology or regulated environments.
• Familiarity with standards like HL7/FHIR or insurance workflows.
• Awareness of compliance requirements (HIPAA, SOC 2).
• Interest in AI safety, reliability, and alignment.

Who You Are

• You thrive in ambiguity and enjoy experimenting your way to clarity.
• You want to see AI solve real, human problems in healthcare, not just toy use cases.
• You’re comfortable moving between roles — one day coding, the next mapping out workflows with clinicians.
• You believe speed, learning, and responsibility go hand in hand.

About the Role

We’re looking for a Prompt Engineer to join our early-stage healthcare AI startup. In this role, you’ll design, test, and optimize prompts and workflows that power real-world use cases such as patient scheduling, insurance eligibility and prior auth, medical history collection, and patient education. You’ll also contribute to model fine-tuning and evaluation to ensure outputs meet the accuracy, compliance, and usability standards needed in healthcare. Expect to wear many hats — from prototyping prompts to coding integrations — as we rapidly iterate and scale.

Key Responsibilities

• Design, refine, and test prompts to enable accurate and reliable outputs for healthcare scheduling, insurance workflows, patient intake, and education.
• Build and maintain reusable prompt libraries and evaluation frameworks to improve performance and reproducibility.
• Contribute to fine-tuning models (collecting and curating datasets, running training jobs, evaluating outputs) to improve domain-specific accuracy.
• Write code to automate testing, integrate AI workflows into APIs and applications, and monitor performance in production.
• Collaborate directly with clinicians and operations teams to understand real-world pain points and rapidly prototype solutions.
• Establish quality guardrails to ensure safety, compliance, and trust in patient-facing outputs.
• Contribute to broader product development beyond prompts—helping with data pipelines, integrations, and deployment.

Required Qualifications

• Bachelor’s degree in Computer Science, Data Science, or related field (or equivalent startup experience).
• Proficiency in Python and/or JavaScript/TypeScript for scripting, automation, and API integration.
• Hands-on experience with large language models (OpenAI, Anthropic, etc.) and prompt design.
• Familiarity with fine-tuning workflows (data preprocessing, evaluation metrics, supervised/LoRA training, or similar).
• Strong problem-solving skills and creativity in working with language and workflows.
• Excellent communication skills to translate between technical, clinical, and operational needs.
• Comfort operating in a fast-paced startup environment with limited structure.

Preferred Qualifications

• Experience with healthcare data standards and workflows (EHR, HL7/FHIR, insurance claims).
• Knowledge of cloud platforms (AWS, GCP, Azure) and deployment best practices.
• Familiarity with compliance frameworks (HIPAA, SOC 2).
• Experience working cross-functionally in small teams where responsibilities overlap.

What We’re Looking For

• Entrepreneurial mindset — resourceful, adaptable, and eager to experiment.
• Excitement about applying AI to reduce friction in healthcare access and operations.
• Willingness to roll up your sleeves and contribute wherever needed — from prompt design to coding to dataset prep for fine-tuning.

We’re building technology that allows patients to smoothly transfer, with all their data, from the hospitalization to the care of their primary providers. There is incredible room and need for technically savvy and creative mind to work hands-on, ground-up to decide how to solve these problems. This involves: Building the infrastructure for hospitals to seamlessly refer patients to Simbie Health, all pending labs/procedures to be communicated, and for Simbie to communicate back updates on the patient. Using LLMs to innovate above and beyond the above basic infrastructure: automatically generating a structured discharge summary from the hospitalization chart, risk stratifying the patient based on their chart and “flagging” the red flag risks to prioritize provider intervention, and completing medication reconciliation with patient education. Integrating with community primary care providers to communicate this info at and after discharge. Making a searchable chronological patient record that can retrieve specific info on prompting from providers (from nurse notes, faxes, scans etc).

We're building with React/Typescript + PostgreSQL.