About Casino Vibe

At Casino Vibe, we create entertainment for players who love casino gaming. As a B2C company, we offer a broad portfolio of casino games, including slots, table games, and live casino, designed to deliver a fun, fair, and immersive experience.

Our team is passionate about technology, player experience, and responsible gaming. We work in a fast-paced, international environment where creativity and collaboration drive our products forward. We are committed to operating with integrity, protecting our players, and promoting safe and responsible play.

About the Role

We're looking for an AI Engineer specialized in building production-grade agentic systems on Claude Code and the Claude Agent SDK. This is a builder role, not a prompting role: you'll design multi-agent architectures, custom tools, and deterministic guardrails that turn LLMs into reliable automation for real business workflows.

Responsibilities

Design and orchestrate multi-agent architectures: decompose complex workflows into isolated subagents (explore, verify, write), each with its own context, reporting only conclusions to the orchestrator
Parallelize agent work while protecting the main agent's context window from pollution
Design and build custom tools (MCP servers, functions, APIs): define what to expose, at what granularity, with model-readable descriptions and context-efficient outputs
Iterate on tool ergonomics based on observed agent behavior — diagnose why an agent fails or loops, and fix it at the tool/architecture level
Implement context management strategies: summarization, delegation, persistent memory (CLAUDE.md, knowledge files)
Build deterministic guardrails with hooks (pre/post-command validation, logging, output checks)
Script the glue: automation in Python/Bash/Node around agentic pipelines

Must-Have Requirements

Hands-on experience with Claude Code and/or Claude Agent SDK in real projects (not tutorials)
Proven tool design experience: you've built MCP servers or custom tool integrations and can show them
Demonstrated subagent/multi-agent orchestration: you can explain how you split tasks, isolated contexts, and merged results
Strong scripting skills: Python, Bash, Node
Solid software engineering fundamentals: Git, testing, CI/CD, API design
Fluent English (written and spoken)
Based in Malta or Europe, Fully Remote

Nice to Have

Published MCP servers, Claude Code plugins, or skills (public repos are your CV)
Experience with hooks and deterministic guardrails around LLM systems
Familiarity with evals and benchmarking of agent performance
Experience integrating agents with business tools (Slack, CRM, ticketing, databases)

How We'll Evaluate You

Live exercise: a real automation task in Claude Code (1–2h, screen share) — architecture decisions matter more than speed.
Deep-dive interview: walk us through an agent you built — what tools you gave it, how you managed context on long tasks, what guardrails you set.

AI Engineer — Agentic Systems (Claude Code)

Job Description