Technical Project Lead, Data Privacy

Protege builds a secure, privacy-centric platform for exchanging AI training data. The Technical Project Lead, Data Privacy leads healthcare data privacy engagements from dataset assessment through validated, certified delivery, and improves the privacy methodology across engagements.

Responsibilities

• 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

Preferred

• 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
• HIPAA De-identification
• Privacy Data
• Healthcare Data
• Named Entity Recognition (NER)
• Optical Character Recognition (OCR)
• Clinical NLP
• Document Processing
• Annotation Projects
• Data Labeling Projects
• Stakeholder Collaboration

Company

• Secure platform for exchanging proprietary AI training data
• New City, New York, USA
• https://www.withprotege.ai/
• Founded in 2024
• 51-200 employees

Qualifications

• 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)
• 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
• Minimum 2 years in data science, applied statistics, ML engineering, or quantitative research, working with real, messy data.

Required skills

• Applied Statistics, Clinical Natural Language Processing, Data Annotation, Data Science, HIPAA De-identification, Healthcare Data Privacy, Language Models, Machine Learning Engineering, Named Entity Recognition (NER), Optical Character Recognition (OCR), Quantitative Research, Statistics

Preferred skills

• $65M, 2024-09-10, 2025-08-13, 2026-01-07, 2026-06-14, 2026-06-20, Andreessen Horowitz, BigDATAwire, Bobby Samuels, Business Ghana, CEO, CRV, Chief Scientific Officer and Co-Founder, Engy Ziedan, Footwork, Growth Stage, Key Investors, Leadership Team, NER, OCR, clinical NLP or document processing, New City, New York, USA, Pulse 2.0, Recent News, Running annotation or labeling projects, Secure platform for exchanging proprietary AI training data., Seed· $10M, Series A· $25M, Series A· $30M, Total Funding, Working with auditors, regulators or external reviewers, https://www.withprotege.ai

Experience

• 2+ years