Experience building and operating a production privacy or data governance system.
Proficient in Python or a similar language, with a detailed understanding of production code.
Direct experience enforcing privacy in AI/ML data pipelines.
Ability to implement pragmatic controls over ideal solutions in real systems.
Experience translating legal obligations into technical requirements.
Responsibilities
Implement and enforce privacy and data governance systems daily.
Build isolation guarantees for customer and user data protection.
Conduct privacy reviews and threat models on new data sources and features.
Translate legal and policy requirements into operational code.
Maintain libraries and guardrails to ensure privacy is the default.
Benefits
Generous health, dental, and vision benefits.
Unlimited PTO.
Paid parental leave.
Relocation support as needed.
Full Job Description
About the Role
We're looking for an engineer to help manage how data is handled across Thinking Machines: what we collect and acquire, where it is stored, who and what systems can access it, how long we keep it, and how it's used in training, evaluation, and products. You'll help build the systems that let us protect privacy and make definitive statements about our practices from data inventory and classification through access and purpose controls to retention, deletion, and audit. You'll also collaborate closely with researchers and product engineers to answer privacy questions that don't have textbook answers.
This is a role for someone who knows privacy deeply and treats it as an engineering problem rather than a checklist. The questions we face are rarely settled by policy alone: what it means to honor a deletion request against a trained model; how customers' fine-tuning data on Tinker should be isolated, never commingled, and only used for the customer; when to deidentify data (and how) vs. apply differential privacy to it. We want someone who can weigh privacy against model quality, research velocity, and product usefulness, pick a defensible answer, and then build it.
What You'll Do
Implement and enforce privacy and data governance across internal systems day to day by building systems like data classification tagging, retention jobs, deletion pipelines, and data classifiers and integrating them across systems at Thinking Machines.
Build the isolation guarantees that keep customer and user data separate and protected, including the training data customers bring to Tinker and the conversations people have with our interfaces.
Run privacy reviews and threat models on new data sources, training runs, and product features, and close them out by shipping the control, the check, or the tooling, not only a recommendation.
Translate legal and policy requirements (GDPR, CCPA, contractual commitments) into working code: policy-as-code, CI checks, and pipeline enforcement, built and maintained by you.
Maintain shared libraries and guardrails that make the private path the default, and fix them when they break or fall out of date.
Skills and Qualifications Minimum qualifications
Built and shipped at least one production privacy or data governance system (retention/deletion pipeline, policy evaluation engine, data cataloging and/or lineage system, consent system, or similar) and owned it through to operation, not just design.
Written and operated production code in Python or a comparable language, evidenced by a system you can walk through in detail: what it does, how it fails, how you debugged it.
Direct experience with enforcing privacy or governing data in an AI/ML data pipeline (training data, fine-tuning, inference, or evaluation) and can describe a specific privacy risk you identified or mitigated in one of these systems.
Has chosen pragmatic controls over ideal ones on real systems, and can describe what was shipped as a result.
Has written a technical requirement or control directly from a legal or policy obligation (GDPR, CCPA, or similar), and can name the obligation and the control.
Preferred qualifications
Experience at an AI lab or with ML infrastructure: training data curation, multi-tenant fine-tuning platforms, or inference systems that handle user data.
Has implemented a privacy-enhancing technique (deidentification, differential privacy, secure enclaves) in a real system, and can describe where it fell short.
Experience building data lineage, metadata, or policy-engine tooling, or large-scale deletion and retention systems.
Background in security engineering, or has worked directly with a security team on access control implementation.
Has been one of the first engineers building out a privacy function, and can point to the specific systems they personally built, not just the program they ran.
Logistics
Location: This role is based in San Francisco, California.
Compensation: Depending on background, skills and experience, the expected annual salary range for this position is $300,000 - $350,000 USD.
Visa sponsorship: We sponsor visas. While we can't guarantee success for every candidate or role, if you're the right fit, we're committed to working through the visa process together.
Benefits: Thinking Machines offers generous health, dental, and vision benefits, unlimited PTO, paid parental leave, and relocation support as needed.
