Trent AI Security Engineer

Trent AI is a London-based agentic security company founded in 2025 by Eno Thereska, Neil Lawrence and Zhenwen Dai, former engineering leaders at Amazon/AWS and Spotify with research backgrounds. It focuses on securing AI agents and the code and infrastructure around them.

The platform uses several specialized AI agents that continuously scan environments, assess risk, mitigate vulnerabilities and evaluate overall security. It works in three steps: understand code, infrastructure, configurations and agents; plan rules mapped to frameworks; and secure by finding risk and fixing it with the customer's coding agents. Design partners include Canopy and Weblogic.

• Understand: reads code, infrastructure, configurations and agents to build one picture of risk.
• Plan: writes security rules with the team, mapped to frameworks.
• Secure: finds risk and fixes it with the customer's coding agents.
• Attack-chain analysis: shows how weaknesses connect.
• Verified fixes: confirms remediation worked.
• Audit trail: records every action taken.

A software company securing code produced with AI coding assistants. A fintech reviewing infrastructure and agent configurations for exploitable attack chains. An IT outsourcing firm adding automated security review to delivery pipelines.

Analyzes attack chains instead of producing long alert lists.

Applies fixes through the customer's own coding agents, with verification.

Keeps a full audit trail of actions.

Operates autonomously or with a human in the loop.

Enterprises are deploying AI agents faster than governance matures; the launch coverage cites a Deloitte survey where 74% of companies plan agentic AI deployment within two years but only 21% report mature governance. Vietnamese software teams adopting AI coding agents face the same security gap.

Founded 2025 in London. 13 million USD seed announced 2026-04-07, led by LocalGlobe and Cambridge Innovation Capital, with executives from OpenAI, Spotify, Databricks and AWS participating. Early access via request or demo; Trust Center linked on the website.