Data Platform Engineer — AI Training Data Ingestion

Dex; Greater London; Hybrid; GBP 90,000 - 130,000; Full time

Job summary

Dex is building a seed-stage research lab in London and Paris, focused on training frontier AI agents. This role owns the data platform powering these environments end to end, from architectural design to production debugging, working across software, platform, and AI engineering boundaries.

You will build high-throughput ingestion, automated PII sanitisation, and validation engines, staying close to production code.

Qualifications

• Proven track record designing, shipping, and operating substantial production data systems end to end.
• Strong software engineering fundamentals; able to debug complex, distributed failures.
• Experience with large, messy datasets and data ingestion/validation/sanitisation challenges.
• Ability to take ambiguous requirements and ship production systems without detailed specs.
• Based in or able to work hybrid from London or Paris.

Responsibilities

• Architect and operate scalable ingestion, processing, and storage systems for high-volume datasets.
• Build automated validation, sanitisation, and PII-redaction pipelines for AI training data.
• Make pragmatic trade-offs across storage, latency, observability, reliability, and cost.
• Establish reusable data platform patterns for other engineers to leverage.
• Debug production failures and deploy modern tooling to keep feedback loops tight.

Skills

Data platform design

Production data systems

Distributed systems debugging

Data ingestion pipelines

Data validation/sanitisation

Cloud architecture

Hybrid work (London/Paris)

Job description

Dex is building a seed-stage research lab in London and Paris, focused on training frontier AI agents. This role owns the data platform powering these environments end to end, from architectural design to production debugging, working across software, platform, and AI engineering boundaries.

You will build high-throughput ingestion, automated PII sanitisation, and validation engines, staying close to production code.