Senior AI Engineer: Build Production-Grade AI Agents

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

Job summary

Dex partners with frontier AI labs to create immersive testing environments and production-grade AI systems. You will own challenging AI and research-engineering problems end-to-end, from data generation to model deployment, including evaluation loops and interactive software integration.

This role emphasizes turning ambiguous questions into robust, observable solutions. We seek hands-on experience with SFT, RL, RLHF or related post-training methods, plus strong Python skills and a track record

Qualifications

• Experience fine-tuning or post-training models and articulating methods and evaluation.
• Strong command of modern AI/ML fundamentals including SFT, RL, RLHF or similar post-training methods.
• Proficiency writing production-grade Python and solving complex engineering challenges.

Responsibilities

• Design methods for generating high-quality experiential data for AI agent training.
• Fine-tune and post-train large models using supervised fine-tuning, RL, and automated optimization techniques.
• Build robust evaluation loops to measure and improve model and agent performance in interactive environments.
• Develop agent harnesses and sandboxed environments connecting AI models to complex software tasks.
• Take ambiguous research questions from concept through to production-grade observable AI systems.

Skills

Fine-tuning / post-training models

SFT / RL / RLHF

Production-grade Python

Ambiguity tolerance / first principles

End-to-end ownership of outcomes

Tools

Python tooling

Job description

Dex partners with frontier AI labs to create immersive testing environments and production-grade AI systems. You will own challenging AI and research-engineering problems end-to-end, from data generation to model deployment, including evaluation loops and interactive software integration.

This role emphasizes turning ambiguous questions into robust, observable solutions. We seek hands-on experience with SFT, RL, RLHF or related post-training methods, plus strong Python skills and a track record