Principal AI Engineer, AI Platform

This role is with one of Dex's trusted partner companies. We work closely with their teams to truly understand their culture, goals, and what they're looking for, so we can match you with the right opportunity and give you context about the role before you commit to a process.

This lab builds the foundational infrastructure for AI agents to learn. They create interactive, resettable digital environments where agents navigate software, complete tasks, and improve through trial and error. This generates the experiential data needed to train increasingly capable AI systems. They're backed by significant funding and work directly with leading frontier AI labs as paying customers.

This is a Staff or Principal-level individual contributor role. You'll own major AI and research-engineering areas, from problem definition and technical direction to measurable production impact. This isn't a management track; it's for engineers who want genuine technical authority and deep hands-on work. You'll turn ambiguous company problems into clear research questions, roadmaps, and success metrics, influencing other experienced engineers through technical credibility.

• Define and lead technical direction for core AI platform areas, from ambiguous problems to clear, measurable outcomes.
• Design and implement systems for training-data creation, curation, model fine-tuning, and post-training optimization.
• Develop and improve agent harnesses, sandboxed execution environments, and robust feedback loops for continuous learning.
• Rigorously evaluate model interventions and agent improvements, establishing baselines and deciding when to iterate, scale, or pivot.
• Influence technical decisions across the team, providing guidance and expertise while remaining deeply embedded in code and experimentation.

• Deep hands-on experience training, fine-tuning, or post-training models, with a track record of rigorously evaluating their impact.
• Strong understanding of modern AI/ML, model optimization, evaluation, and reinforcement learning techniques, coupled with production-quality software engineering skills.
• Proven ability to translate highly ambiguous technical or research problems into clear plans and successful, long-term initiatives.
• Demonstrated leadership through technical credibility, defining goals and direction for multiple contributors without needing line-management authority.
• Willingness to work from London or Paris, where the engineering hubs are located.