Agentic AI Systems Research Engineer

Callosum; Greater London; On-site; GBP 100,000 - 150,000; Full time

Benefits offered by this job

Private healthcare

Visa sponsorship

Relocation assistance

Job summary

Callosum is building the infrastructure for autonomous, heterogeneous AI systems in London. The role focuses on designing and running experiments to understand why agentic AI systems fail and how to improve them at scale, across memory, task decomposition, and inter-agent coordination.

You will work with cutting-edge models and silicon, contribute to durable evaluation tools, and publish findings to advance the field. London office-based, visa sponsorship available for top talent.

Qualifications

• Evidence you can run your own research: design experiments and publish results.
• Hands-on experience with LLMs in agentic settings and tool use.
• Discipline in experiments: hypotheses, controls, ablations, and clear results.
• Strong engineering: you write code to run your experiments.
• Clear, prioritised written guidance for teams and external audiences.

Responsibilities

• Design and run experiments isolating why agentic systems fail across behaviour and activations.
• Develop infrastructure for analysis, observability, and evaluation.
• Publish evaluation results and methods as company assets.
• Collaborate across teams to deploy scalable, cost-efficient solutions.

Skills

Research design

LLMs in agentic settings

Engineering experiments

Strong communication

Tools

SGLang

vLLM

Inference serving

Job description

Callosum is building the infrastructure for autonomous, heterogeneous AI systems in London. The role focuses on designing and running experiments to understand why agentic AI systems fail and how to improve them at scale, across memory, task decomposition, and inter-agent coordination.

You will work with cutting-edge models and silicon, contribute to durable evaluation tools, and publish findings to advance the field. London office-based, visa sponsorship available for top talent.