Software Engineer

Phantom Studios builds software for forensic accounting, KYC and compliance, and fraud work, including our own products PhantomTrace and our onboarding product. We may also take on software work for eGaming companies. This is a small team, so what you build ships and gets used.

You'll report to the head of engineering, who leads our software and AI work. You will own projects, not tickets. You take something from a rough idea to a running thing people rely on, and you keep it running.

We build with AI agents every day. Claude Code is central to how we work, so this role sits right in the middle of that.

What You'll do

• Build and own some of our projects end to end: backend, deployment, and the day-to-day of keeping them healthy.
• Design systems that are simple to run and simple for the next person to read.
• Work with AI agents as a normal part of how you build, and know when to trust them and when not to.
• Integrate with third-party services (KYC providers, payment and gaming platforms, forensic data sources).
• Pick up whatever the project needs, including bits of frontend, data work, and infrastructure.

What you need

• Strong backend engineering in at least one language, and the sense to pick the right tool rather than the familiar one.
• Cloud and deployment. You can put a service into production and look after it, not just write the code.
• Solid grounding in system design and engineering principles. You think about failure, data, and maintenance before you think about frameworks.
• Fluent with agentic development. You've built real things with coding agents (Claude Code or similar) and you understand context management, skills and subagents, prompting, and orchestration. You have a view on where they help and where they get in the way.
• API design and integration. Clean interfaces, sensible contracts, third-party services wired in without drama.

Nice to have

• Enough frontend and UI sense to build a decent interface, even if you're not a frontend specialist.
• Python.
• Application security, and care when handling sensitive data.
• Any exposure to forensics, compliance, fraud, or eGaming. Not required, we'll teach the domain.

How we work

Claude Code and agents like it run through things we build, with custom skills, subagents, and careful context management. We're not hiding that, and you shouldn't either. We don't care whether you already know our exact setup. We care whether you learn fast and whether your judgment holds up when the agent is confidently wrong. Two things matter most to us:

• Learning velocity. You pick up new tools, domains, and codebases quickly and turn them into working software.
• Judgment with AI output. You verify before you ship. You know when to drive the agent, when to take over, and when to write it yourself.

The Process

• A conversation where you walk us through something you've built and the decisions behind it.
• An in-office pairing session on the Isle of Man: a real repo, a real task, and the tools you'd actually use, including AI agents. We watch how you work, not whether you memorised anything.

Practical

• Full-time employee, based on the Isle of Man.
• You must already have the right to live and work on the Isle of Man. If you'd need a work permit, we can look at it case by case, but we can't guarantee sponsorship.