Member of Technical Staff, AI (up to £200k)

Dex; Greater London; On-site; GBP 120,000 - 180,000; Full time

Job summary

Dex is building production-ready autonomous agent technology in London, focusing on end-to-end agent architecture, safety guardrails, and closed-loop evaluation. You will own harness design, from memory management to multi-agent orchestration, working closely with founders who value unconventional technical opinions and fast execution.

The role emphasizes deploying agents in real enterprise environments, shaping the technical direction, and contributing to foundational systems engineering beyond

Qualifications

• Hands-on experience building and deploying autonomous agents in production.
• Harness engineering depth across orchestration, tool use, memory systems and eval pipelines.
• Strong ML foundations or senior applied AI with full-stack execution.
• Preference for guardrails around model indeterminism.
• Based in or able to work regularly from the London King's Cross R&D hub.

Responsibilities

• Architect autonomous agent harnesses capable of executing complex workflows in enterprise environments.
• Design safety guardrails and runtime constraints for model indeterminism.
• Build closed-loop evaluation systems to benchmark and improve harness performance.
• Ship agent features directly into production with enterprise partners.
• Contribute to foundational technical strategy from memory management to multi-agent orchestration.

Skills

Autonomous agents

Orchestration

Memory systems

Evaluation pipelines

Full-stack execution

Job description

The role

Most enterprise AI stalls because models cannot be trusted to act autonomously inside production systems. This stealth team is solving the harness and safety layer directly, backed by founders who previously scaled enterprise AI and agentic security products from zero to eight figures. With 11 people, secure early funding, and active enterprise design partners, they are building production infrastructure rather than research demos, with a founder culture that commits to backing its own alumni when they eventually launch companies. As one of two founding Member of Technical Staff seats in London, you will own applied AI and agent architecture end to end. This is not prompt wrapping or chaining basic API calls; it is core systems engineering around model indeterminism, autonomous tool use, and closed-loop evaluation. You will set technical direction with founders who want strong, unconventional technical opinions and zero corporate process.

The work

• Architect autonomous agent harnesses capable of executing complex workflows inside enterprise environments.
• Design safety guardrails and runtime constraints that manage model indeterminism without crippling agent utility.
• Build closed-loop evaluation systems that continuously benchmark and systematically improve harness performance.
• Ship agent features directly into production with large enterprise design partners.
• Contribute to foundational technical strategy, from memory management to multi-agent orchestration.

What You Bring

• At least one year of hands-on experience building and deploying autonomous agents into real production environments, beyond prototyping with LLM APIs.
• Demonstrated harness engineering depth across orchestration, tool use, memory systems, and systematic eval pipelines.
• Either strong ML research foundations applied to agent architectures or senior applied AI experience paired with full-stack execution.
• A clear preference for engineering guardrails around model indeterminism rather than stripping agents back to rigid, deterministic rules.
• Based in or able to work regularly from the London King's Cross R&D hub.

As part of the recruitment process at Dex, we process your personal data in accordance with our Privacy Notice for Job Applicants. This notice explains how and why your data is collected and used, and how you can contact us if you have any concerns.