Want to know if this job is worth applying to?
New York, New York, United States
Onsite
Not Available
Company Overview:
We are building Protege to solve the biggest unmet need in AI — getting access to the right training data. The process today is time intensive, incredibly expensive, and often ends in failure. The Protege platform facilitates the secure, efficient, and privacy-centric exchange of AI training data.
Solving AI’s data problem is a generational opportunity. We’re backed by world-class investors and already powering partnerships with some of the most ambitious teams in AI. The company that succeeds will be one of the largest in AI — and in tech.
We’re a lean, fast-moving, high-trust team of builders who are obsessed with velocity and impact. Our culture is built for people who thrive on ambiguity, own outcomes, and want to shape the future of data and AI.
Healthcare data only becomes AI training data once patient privacy is protected to a standard an independent expert will certify, while keeping the clinical detail that makes it worth training on. Black out too little and patients are exposed. Black out too much and the data loses its value.
No single tool gets this right, because every dataset is different. A radiology report, a handwritten prior-authorization fax and a veterinarian's notes each fail in their own ways. So we've built a practice instead of a product. It combines off-the-shelf tools, custom rules, language models and human review, tunes the mix for each dataset, and proves the result with the kind of statistical validation an auditor would accept. It has cleared every recent independent review with zero failures.
You'll run it. You'll lead engagements from the first look at a new dataset to certified delivery, make the calls on what's safe to ship, and improve the method with every engagement.
Lead data privacy projects from first look to certified delivery: assess the data, write the policy, tune the pipeline, validate, and deliver.
Design and run the experiments that decide which layers a dataset needs, measuring recall, over-redaction and cost for each.
Run validation the way an auditor would: set the pass bar before you look, test on fresh held-out samples, report confidence bounds, and never fix the test set.
Turn every miss into a lasting fix (a rule, a model or prompt change, or a human-review step) and lock it in with a regression test.
Work directly with independent expert determiners: prepare evidence, answer methodology questions, and respond to findings.
Direct human reviewers on the hard cases and adjudicate their disagreements.
Evaluate new tools and models (open-source NER, vendor APIs, models that run inside a partner's own environment) against our benchmarks.
Write down what you learn so the next engagement starts ahead of the last one.
30 days: learn the method. You know the playbook and tooling, you've shadowed a live engagement, and you've delivered your first intake assessment of a new partner's data.
60 days: share the load. You've co-led an engagement through validation, owned the experiment that set its pipeline, and added new rules with regression tests.
90 days: own it. You've led an engagement end to end, including the evidence package for the expert reviewer, and you've evaluated a candidate model with a clear keep-or-drop recommendation.
Must have
Minimum 2 years in data science, applied statistics, ML engineering, or quantitative research, working with real, messy data.
Solid statistics: sampling, confidence intervals, measurement design and inter-rater agreement. You can explain why a sample is big enough, and what it can't tell you.
Strong background, and experience using language models or AI coding tools to move fast.
Healthy skepticism. You verify claims before trusting them, including your own.
Clear writing for technical and non-technical readers, including external reviewers.
Authorization to work in the U.S. as a W2 employee (required for access to protected health information).
Nice to have
HIPAA de-identification experience (Safe Harbor or expert determination), or other privacy or healthcare data work
NER, OCR, clinical NLP or document processing
Working with auditors, regulators or external reviewers
Running annotation or labeling projects
Pass the Loved Ones' Test. We act with integrity and do the right thing — especially when it's hard and no one is watching.
Always Find a Way. We are resourceful, resilient builders who solve hard problems and push through obstacles.
Go Fast and Grow Fast. Velocity matters. We move with urgency, learn quickly, and continuously improve as individuals and as a company.
Practice Kindness and Candor. We communicate directly and respectfully, building trust through honest feedback and genuine care for one another.
Deliver Together. We win as one team. Collaboration, accountability, and shared ownership drive our success.
Own the Outcome. Hone the Craft. We take pride in our work, sweat the details, and continuously raise the bar for excellence.
