About Healthleap

HealthLeap builds AI that helps clinicians prioritize patients, surfaces the right data, and gets patients the care they need earlier, so they can leave the hospital sooner.

We integrate with hospital electronic health record systems, screen 100% of patients daily, and risk-rank them in real time. Clinicians at Cedars-Sinai and Penn Medicine start every morning with HealthLeap — with Houston Methodist, Emory, and Intermountain Health deploying now.

Real results: 39% more diagnoses. 4 days earlier detection. $11M/year ROI for our first site at Cedars Sinai. 7× revenue growth in 7 months.

We started with malnutrition. We're expanding to every major condition to ensure no patient falls through the cracks. Sequoia and First Round are backing us to build the platform that screens every patient for everything and drives tangible outcomes.

We're ~15 people. >$7M raised. SF-based, hybrid-friendly. Early enough to shape the product. Late enough to know it works. Results that are changing lives.

About the Role

Build the models that tell clinicians which patients need attention.

HealthLeap's ML models run every morning at Cedars-Sinai, Penn Medicine, and Houston Methodist - surfacing the patients who would otherwise slip through the cracks. You'll improve these models, build feedback loops that actually learn, and expand to new conditions that affect millions of patients.

Why you

You've taken models from notebook to production and kept them running. You've worked with messy real-world tabular data, not clean academic datasets. You care about impact - models that help patients, not papers. You're comfortable with the unpredictability of startups. You're genuinely excited about healthcare.

What you'll do

Improve our malnutrition model - better feedback loops, fewer false positives
Build features from structured EHR data - labs, vitals, patient history
Expand to new conditions - CHF, pressure injury, COPD, and beyond
Own the full lifecycle: train → deploy → monitor → retrain
Work with backend to receive data from pipelines and make it useful for ML
Talk to clinicians and customers to understand what models should actually do

What we're looking for

3-5+ years ML engineering experience in a high-growth environment
Strong production ML skills - you've deployed models, not just trained them
Deep experience with tabular data, feature engineering, classical ML
Understand model explainability - clinicians ask "why did this patient score high?"
Genuinely excited about healthcare and patient impact
Comfortable with unpredictability - priorities shift, you adapt

Nice to have

Healthcare or clinical ML experience
Background at an applied AI company with real product traction
Experience building feedback loops and continuous model improvement
Side projects that show you can productionize, not just prototype

This role is NOT for you if

Startup unpredictability feels like chaos to you. We find it exciting.
You need detailed specs before you start. We figure it out as we go.
You've never owned something end-to-end. Here, you own outcomes, not tasks.
You wait to be told what to do next. We need people who see what's needed and do it.
You're looking for a 9-5 with predictable hours. We care about deep work and deep rest (we have a minimum leave policy), but when a hospital go-live is on the line, we show up - even if that means a 60+ hour week.

Interview process

Intro call - Get to know each other
Technical - 1-2 interviews on ML system design + past work deep dive
Onsite - Coding, case study, team meet
Decision - Same week as onsite

We respect your time. If there's a fit, you'll know fast.

Compensation & Benefits

Salary: $200,000 - $325,000 base
Equity: Meaningful ownership in an early-stage company
Healthcare: 100% of premiums covered
PTO: Unlimited, with a recommended minimum of 20 days
401(k): 4% match
Equipment: Laptop + budget for your home office

Location

San Francisco, hybrid (Mon/Wed/Fri in office)

If you're passionate about applying frontier AI to real-world impact, join us in building healthcare's future.