Job Description

Are you looking to work in the cutting edge area of applying data-science to help global customers get a better insight into their health? If so, read on and apply.

n

n

Role Name: Senior Data Scientist

n

Science Team | Full-Time | In-Office | Bangalore (Kudlu Gate)

n

Salary: ~30L-40L+ESOPs

n

n

The Role

n

The Ultrahuman Science Team builds the algorithms behind the Ring, M1 CGM, blood and urine biomarkers, and Performance Lab assessments. We are hiring a Senior Data Scientist to own those algorithms end to end: from the raw sensor signal to a model that is shipped, monitored, and trusted in users' hands.

n

This is a builder role with real scope. In a typical month you will improve a production algorithm, root-cause a metric users are complaining about, and stand up the data pipeline the next model needs. The common thread is ownership: you take a vague question and return a working answer, without waiting to be handed scope.

n

n

What You'll Do

n

n
- Own algorithms end to end: sleep staging, activity detection, sensor-derived metrics, and health scores. You frame the problem, build the features, train and evaluate the model, and see it live
n
- Ship models, not notebooks: you prove a change on our own cohort before it reaches users, and a model is done only when it runs in production and you can tell how it is behaving
n
- Validate against reference standards: design evaluations against gold standards, reference devices, and study ground truth, and know when a result is real and when it is an artifact
n
- Own the data layer: cohort extraction, feature pipelines, study data, and raw sensor data, so the next model starts from clean inputs
n

n

n

What This Looks Like in Practice

n

1. Improving production algorithms - Take an existing production model like sleep staging, root-cause the failure modes against reference data, and ship a fix you can defend with numbers.

n

2.



Building new models - Train an activity classifier on raw sensor data, design the labeled data collection that expands it, and pick the operating point so false positives never erode trust.

n

3. Proving it before it ships - Run a new steps algorithm against reference-device cohorts, decide with data when it is ready, and monitor how it behaves after rollout.

n

n

Who You Are

n

The two things we can't coach

n

n
- High ownership, end to end: you take a problem from a vague question to a shipped model without waiting to be handed scope, and you can point to something you owned from raw data all the way to production
n
- Hungry for more scope: you have outgrown your current role and want problems biggerthan your title, with the technical depth to be trusted with them
n

n

n

Also important

n

n
- You've worked with human health data: wearables, physiological signals, or clinical data.
n

n

n

If your experience is close but not exact, show us why you will ramp rapid

n

n
- You've built at a startup: or somewhere small enough that nobody handed you clean data, clear specs, or a mature ML platform
n
- You work like it's 2026: coding agents and AI tooling are part of how you build every day, and you can tell which new capabilities are worth adopting
n
- You communicate: you can explain a model and its limits to a product manager, an engineer, or a founder, and hold your own with our scientists Core Technical Skills
n
- Languages and data: Python and SQL daily, comfortable working in a real codebase
n
- Machine learning: PyTorch or TensorFlow, scikit-learn, and gradient boosting,



with the judgment to know which the problem needs
n
- Advanced machine learning: time series and sequence models, deep learning on continuous physiological signals, and ensembles
n
- Statistics and evaluation: hypothesis testing, experiment and A/B design, model evaluation, and error analysis against a reference standard
n
- Scale and cloud: Spark or equivalent on large datasets, and AWS, GCP, or Azure
n
- Production ML and MLOps: training pipelines, model versioning, deployment, monitoring, and drift detection
n
- LLMs and agentic systems: fine-tuning and serving models, building agentic pipelines, and using coding agents to move faster
n

n

n

Experience:

n

- 4 to 5 years building and shipping machine learning systems. We index on what you have shipped and on trajectory, not the exact number of years; if you are a little earlier but have clearly outgrown your current scope, we want to hear from you.

n

- Bachelor's or higher in engineering, computer science, statistics, or a related field.

n

n

How We Work and Who Thrives Here

n

- The Science team is small and moves fast, and much of the work has no precedent to copy.

n

- People do their best work here when they are energized by ambiguity, low on ego, quick to adopt a better idea no matter where it comes from, and comfortable owning something before anyone has told them how. If you need a mature data org, clean labelled datasets, and clear guardrails to thrive, this particular role will not be the right fit, and that is worth knowing up front.

n

n

What You'll Gain

n

n
- Ownership of algorithms that hundreds of thousands of people see every morning
n
- A dataset most scientists never get to touch: 100M+ nights of sleep and continuous physiological signals at scale
n
- Direct collaboration with the engineering, product, and design teams building Ultrahuman
n

📌 Senior Data Scientist (Bengaluru)
🏢 Orangemint Technologies
📍 Bengaluru