Beazley Security, LLC

Compensation

Full Description

Summary Beazley Security has Claude and Microsoft Copilot deployed across the company. The next step is putting agents to work across every team in the business. This is our first hire dedicated to making that happen. You’ll work with operations, sales, finance, and other non-technical teams to find manual work worth automating. You’ll build agents in Claude and Copilot Studio, get them into daily use, and measure the time they save. Early on you’ll build most of them yourself, while teaching teams to build their own. You’ll have a big say in what we tackle first and how we build it. This is the first hire for what we plan to grow into a team, so a lot of what you set up early will stick. Beazley Security protects client data and systems for a living, so you’ll build security and privacy into every agent from the start. You’ll help us establish safeguards to ensure that agents can go into production quickly and safely, by building upon our existing agentic security tooling and capabilities. Responsibilities Spend time with teams across the business, watch how the work gets done, and pick the problems where an agent saves hours or removes a manual, error-prone step. Agree on what success looks like before you build. Turn a stream of AI ideas into a short, ranked backlog with a clear call on each: build it, help the team build it, or leave it alone. Estimate what each agent will cost to run, including licensing and token consumption. Use our pre-agreed guardrails so low-risk agents go live quickly, and bring in Security, IT, Legal, or Privacy when a use case needs a closer look. Design, build, test, and deploy AI agents and supporting automation using the Claude Platform and Microsoft Copilot Studio. Build a group of champions across the business, run hands-on build sessions and office hours, and create the templates and playbooks teams use to build their own agents safely. Help maintain a record of every agent in production with its owner, data sources, and access. Evaluate new AI tools when teams ask about them and recommend whether we adopt them. Your First 90 Days Days 1 to 30: Get to know Beazley Security. Sit with at least three teams, map their most painful workflows, and build your first agent yourself. Get it into real use, even if it’s small. Days 31 to 60: Two or three agents in production, each measured against a baseline of “value”. A published intake process, and your first AI champions identified. Days 61 to 90: At least one team can point to manual work that’s gone. The agent register and build playbook are live. After that: Champions build their own agents using your patterns, and your results justify the next hire onto this team. What You'll Have Executive buy-in and support. You report to our Chief Product and Technology Officer, who uses AI every day and has built several agents for internal use. Expect quick decisions and direct help when a project gets stuck. A first project. A business sponsor and workflow lined up before day one. A fast review path. Guardrails agreed with Security and Legal before you start, so low-risk agents don’t sit in a queue. Qualifications You’ve built AI agents or automations that are in daily use, with low-code tools like Copilot Studio or with code (Python, TypeScript, and the Claude API). You have hands-on experience with prompt and context design, retrieval, tool use, and evaluating agent output, and you know where LLMs fail. You’ve integrated with REST APIs, MCP servers, connectors, and enterprise authentication. You’ve typically spent four or more years in a business-facing technical role such as solutions engineering, solutions consulting, business analysis, automation, or platform delivery. You can learn a finance or sales operations workflow from the people who run it and build something that improves it. You know when a use case needs a security or privacy review before it goes anywhere, and you can estimate what an agent will cost to run on Copilot credits or Claude API usage. You’ve run workshops or training for non-technical audiences, and you write clearly. You’d rather ship a small agent that works than constantly plan and iterate on a big one. Preferred Experience in cybersecurity. Experience running an AI champions program, community of practice, or citizen developer program. Experience setting up lightweight AI guardrails, like an agent inventory or risk tiers, that keep teams moving quickly. Familiarity EU data protection requirements. When you apply, please include a link or a short screen recording of something you built. Beazley Security Offers Competitive salary and bonus. Flexible working arrangements. Generous leave policies including 3 months paid parental. 100% of employee-only insurance premiums covered (healthcare, dental and vision). Up to 5% matched 401k contribution. Opportunities for career advancement and ongoing training. Participation in industry conferences and events.