Research Engineer: Agent Systems & RL for Long-Horizon AI

Huawei; City of Westminster; On-site; GBP 55,000 - 90,000; Full time

Job summary

Huawei UKRD London Office seeks researchers in Agents and reinforcement learning to advance long-context Agent memory systems and continuous self-evolution of the Agent Harness framework. You will work on Agentic RL techniques, connecting policy models with practical deployment pipelines and collaborating with Huawei’s UK R&D teams.

Ideal candidates have strong engineering or theoretical RL backgrounds, high-impact publications, and a track record in competitive AI research.

Qualifications

• Solid engineering implementation skills or strong theoretical foundations; excel in at least one.
• In-depth theoretical understanding or hands-on experience in Agents, RL, and memory systems.
• First-author publications at top-tier conferences or journals (ICLR/NeurIPS/ICML/ACL/EMNLP).
• High-citation papers or influential open-source Agent projects.
• Outstanding rankings in relevant competitions (Kaggle, ARC-AGI).

Responsibilities

• Design and develop Agent memory systems for ultra-long context processing.
• Explore unified representation and optimisation techniques for Agent Harness.
• Investigate Agent optimisation based on parametric and non-parametric RL and build update pipelines.
• Collaborate with Huawei Research and Development UK Limited to deliver results.

Skills

Agent systems

Reinforcement learning

Memory systems

Self-evolution

Agent Harness

RL pipelines

Job description

Huawei UKRD London Office seeks researchers in Agents and reinforcement learning to advance long-context Agent memory systems and continuous self-evolution of the Agent Harness framework. You will work on Agentic RL techniques, connecting policy models with practical deployment pipelines and collaborating with Huawei’s UK R&D teams.

Ideal candidates have strong engineering or theoretical RL backgrounds, high-impact publications, and a track record in competitive AI research.