Role: Java Developer (AI-First)
Location: Cape Town City Centre (in-office)
Global Scope: Opportunities and exposure across international teams (UK, China, Philippines and more).
Remuneration: Negotiable for the right candidate
We are assisting our client, an international 5G engineering and design business, source for a Cape Town-based "AI obsessed" Java Developer.
**Please Note: The AI component of this role is very important**
About the role:
We’re sourcing a full-stack developer with deep Java expertise and serious AI tooling skill. This is an AI-first engineering role: our client's stack relies heavily on Claude Code as a core part of how they build, and they expect you to be one of the strongest AI-assisted engineers in the building!\
You’ll ship features end-to-end, from database and Spring Boot services through to the frontend, at a pace that’s only possible when you drive AI tools well.
You’ll report to our client's Senior Java Team Lead and work on high-volume, low-latency systems on a modern Java / Spring Boot stack. You’ll get real ownership of features, direct input on technical decisions, and the tooling budget to work the way modern engineers should.
What you will actually do:
Build, Full Stack
- End-to-End Ownership: Design and build features end-to-end: Spring Boot services, REST APIs, database schemas, and the modern frontend that sits on top.
- Production Java: Write production Java (17+) daily. Own the quality of what you ship, tests included.
- Data & Cloud: Design schemas and optimize queries in PostgreSQL; deploy and run services on AWS with Docker and CI/CD pipelines.
- Team Collaboration: Contribute actively to technical design discussions and code reviews across the team.
Engineer with Claude Code
- Daily Driving: Use Claude Code for scaffolding services, writing and refactoring code, generating test suites, debugging, and exploring unfamiliar parts of the codebase.
- Agentic Workflows: Set up repository context (CLAUDE.md, project conventions, custom instructions) and leverage MCP servers to maximize output quality.
- Fast Spikes to Hardened Code: Vibe-code prototypes and internal tools fast, then know exactly how to harden and refactor them before they touch production.
- Critical Review: Review every line of AI output critically. You catch subtle bugs, edge cases, and security gaps—you never ship code you can’t explain.
- Raise the Bar: Share prompting strategies, Claude workflows, and agent setups to elevate the whole engineering team’s AI game.
What we need from you:
AI tooling: Expert, Fluent, Not Dependent
- Advanced, daily Claude Code use: You can point it at a real codebase and get production-quality results: multi-file changes, refactors, test suites, debugging sessions. If we asked you to demo this in an interview, you’d enjoy it.
- Working with AI coding agents: You understand how to break work into well-scoped tasks, write effective context and instructions (CLAUDE.md files, project conventions), and iterate effectively when output misses.
- Fluent with vibe coding and its limits: Fast prototypes, yes. Unreviewed code in production, never.
- Strong fundamentals: Strong enough fundamentals to work entirely without AI when needed. You can debug, design, and reason about systems on your own. AI multiplies you; it doesn’t carry you.
- Red flags: Engineers who paste in AI output without reading it, or who can’t explain the code they submitted.
Experience:
- 5+ years building production Java systems. Strong Spring Boot and modern Java (17+); you can build a service from scratch and explain every choice.
- Genuine full stack capability: Comfortable building and maintaining frontend code in a modern framework (React, Angular, or Vue with TypeScript preferred).
- Experience with the Claude API, MCP servers, or building AI-powered features into products.
- Microservices experience: Service boundaries, API design, async messaging, observability.
- Solid with PostgreSQL or MySQL: Schema design, query tuning, migrations.
- AWS in production (ECS/EKS, S3, IAM); Docker fluency; CI/CD pipelines you’ve built or meaningfully improved. Terraform is a plus.
- Linux & Git: Comfortable on Linux; Git workflow fluency (branching, code review, GitHub or GitLab).
- Agile / Scrum: You’ve worked in Agile / Scrum teams and deliver reliably inside them.
How you work:
- You explain your thinking clearly, in writing and out loud.
- You move fast without being careless. Speed from tooling and judgement, never from skipped review.
- You give and take direct, kind feedback.
- You finish things. “Almost done” doesn’t count.
Qualifications:
- BSc / Diploma in Computer Science, Engineering, or equivalent practical experience.
- Oracle Certified Professional (Java) is a plus, not a requirement.
Nice to have:
- A public body of AI-assisted work we can look at: projects, tooling, agents, write-ups on your workflows.
- Event-driven architectures (Kafka, RabbitMQ, SQS).
- Performance tuning on low-latency systems (JVM tuning, profiling, caching).
- Exposure to Kotlin, Go, or other JVM / backend languages.
What we offer:
- AI Tooling: An AI-first engineering culture with the tooling budget to match: Claude Code and the stack to use it properly.
- Real Autonomy: A small team, real autonomy, and features you own end to end with a modern developer-led stack and permission to keep it that way.
- In-Office Culture: Collaborative, high-bandwidth environment in Cape Town (this is not a remote role).
- Global Scope: Opportunities and exposure across international teams (UK, China, Philippines and more).