Privacy and Anonymization Research Engineer

Quick Facts

• Role: Privacy and Anonymization Research Engineer
• Work model: On-site in San Francisco or Singapore, or remote with ~70–80% overlap with US Pacific or Singapore business hours
• Start: ASAP

Description

Build privacy-preserving systems that detect and remove PII, secrets, credentials, and other sensitive content from raw real-world data before it enters AI training and synthetic data pipelines. Own an end-to-end pipeline for detection methods, anonymization transformations, and evaluation of privacy risk versus retained data utility. This role focuses on privacy and data-protection engineering ahead of downstream processing—not on synthetic task design.

Responsibilities

• Build systems to detect PII, quasi-identifiers, credentials, and other sensitive information; design type-appropriate transformations.
• Develop and benchmark detection approaches using rules, statistical models, classifiers, and LLM-based methods.
• Build production pipelines that anonymize raw data prior to downstream processing, training, evaluation, and synthetic generation.
• Create evaluation frameworks for privacy risk and utility, including recall-weighted metrics, leakage tests, and adversarial re-identification attempts.
• Ensure robustness to new data sources, schema drift, unusual formats, and sensitive information in unexpected fields.
• Translate privacy requirements into practical technical policies and safeguards by collaborating with engineering, research, operations, and customers.

Requirements

• Strong proficiency in Python; experience building reliable production data or machine-learning systems.
• Experience with information extraction, NER, classification, or related methods for detecting rare or sensitive content.
• Experimental instincts to compare approaches across recall, precision, latency, cost, and downstream data utility.
• Understanding of redaction, masking, pseudonymization, anonymization, and synthetic data and when each is appropriate.
• High attention to detail for reasoning about subtle leakage paths, edge cases, and adversarial failure modes.
• Experience building end-to-end data-processing pipelines, including in situations without a fully prescribed roadmap.
• Clear written and spoken English and comfort collaborating across teams and time zones.

Benefits

• Make real-world data safe for AI training by protecting people and secrets without destroying useful structure and signal.
• Own the full technical path from detection to production: detection methods, transformations, evaluation, and the front-line anonymization pipeline for training and generation.