Data engineer in AI security

Refractal; Greater London; On-site; GBP 80,000 - 135,000; Full time

Benefits offered by this job

Equity

Pension

Private medical and dental cover

Visa and relocation support

Job summary

Refractal, a UK-based AI security company, is hiring a data engineer to build and govern the evidence store powering audits and compliance.

You will own production data pipelines, design robust data models for multi-tenant environments, and collaborate with data scientists and security engineers to create provenance-rich training datasets. The role offers equity, pension, private medical cover, and visa support.

Qualifications

• 6+ years in data engineering with ownership of production pipelines.
• Strong Python and SQL skills, with PostgreSQL experience.
• Experience with streaming and orchestration tools (Kafka, Airflow, Dagster).
• Data modelling for audit trails, lineage and multi-tenant isolation.
• Security-minded handling of sensitive data; familiarity with UK GDPR is a plus.

Responsibilities

• Design the data model for an append-only, cryptographically signed evidence store.
• Build streaming and batch processing to turn ingested activity into queryable data.
• Own data quality, retention, residency, and deletion controls for regulated customers.
• Create training and evaluation datasets with data scientists and security engineers.
• Enable quick compliance reporting against EU AI Act, NIST AI RMF, and ISO/IEC 42001.
• Ensure the evidence store is in production, documented, and trusted by customers' auditors.
• Answer compliance and detection questions from data you own, avoiding ad hoc work.
• Maintain provenance and separation for training/evaluation data.

Skills

Python

SQL

PostgreSQL

Data warehouse

Kafka

Airflow

Dagster

Tools

Kafka

Airflow

Dagster

Job description

Описание:

Refractal is an AI security company building the infrastructure organisations need to defend against autonomous AI threats. 2026 has shown that generally capable intelligence is here. The challenge now is modernising security to stop AI-enabled threats, whether they come from internal agents going rogue or from attackers using AI to become more dangerous. You'll build Refractal's data foundation, including the evidence store that customers' auditors rely on, the analytics behind compliance reporting and the datasets our models are trained on.

Задачи:

• Design the data model for our append-only, cryptographically signed evidence store.
• Build the streaming and batch layer that turns ingested activity and detections into queryable data, working with our integrations engineers on schemas.
• Own data quality, retention, residency and deletion controls for regulated customers.
• Build training and evaluation datasets with our data scientists and security engineers.
• Make compliance reporting against the EU AI Act, NIST AI RMF and ISO/IEC 42001 quick to produce from our data.
• The evidence store is in production, documented and trusted by customers' auditors.
• Colleagues answer compliance and detection questions from data you own, without ad hoc work.
• Training and evaluation data has clear provenance and separation.

Требования:

• 6+ years in data engineering, owning production pipelines.
• Strong Python and SQL, with Postgres plus a warehouse or analytics store such as BigQuery or ClickHouse.
• Experience with streaming (Kafka, Pub/Sub or similar) and orchestration (Airflow, Dagster or similar).
• Data modelling for audit trails, lineage and multi‑tenant isolation.
• Security‑minded handling of sensitive data.
• Будет плюсом: Data work in banking, insurance or other regulated industries; Experience with UK GDPR and data residency

Условия:

• Salary: £80,000–£135,000
• Competitive equity to share in Refractal's growth
• Pension contribution
• Private medical and dental cover
• Visa and relocation support