Python Developer (AI)

We are looking for a solution-oriented AI Engineer who can design and build production-grade multi-agent LLM systems.

The key expectation is not just writing code, but understanding the desired user outcome and engineering systems that reliably deliver it.

This role is closer to systems engineering / architecture than classical ML.

You will work independently, explore our APIs and product flows, and build agent-based solutions that integrate deeply with the platform.

## Context

• Product platform with multiple internal APIs
• Agents interact with platform APIs and product workflows
• High variability of use cases and flows
• Relatively small data volumes per agent, but high orchestration complexity
• Agents must operate autonomously with tools, routing, memory and fallback strategies
• Backend stack mainly Python, frontend TypeScript

## Responsibilities

• Design and implement multi-agent systems (orchestrator + worker agents)
• Build agent infrastructure:
• routing
• memory
• tool execution
• fallback strategies
• Develop multimodal pipelines (text images)
• Integrate and abstract different model providers
• Build custom tools and MCP servers
• Connect agents to internal platform APIs
• Read and analyze platform API specs and source code (Python backend / TypeScript frontend)
• Extend or build API layers / adapters when necessary
• Implement automated testing and LLM evaluation pipelines
• Debug agent behavior and improve reliability

## Requirements

• Familiarity with the gamedev domain, including experience in analytics, payments, or similar areas of the industry
• Strong Python backend development (async, APIs, services)
• Commercial experience building multi-agent LLM systems
• Experience with agent frameworks like:
• LangChain
• LangGraph
• AutoGen
• CrewAI
• LlamaIndex
• Pydantic AI
• Experience designing:
• agent orchestration
• tool calling
• routing
• memory systems
• Experience building multimodal pipelines (text + image generation)
• Experience integrating multiple model providers:
• OpenAI
• Anthropic
• local / self-hosted models
• Vector databases: Qdrant / Pinecone / Weaviate
• Redis or similar caching / state layers
• Experience with production-grade development practices
• testing
• debugging
• observability

### Important: commercial experience is required (not pet projects).

## Nice to have

• ML background or practical ML knowledge
• Experience designing AI product architectures
• Experience building agent evaluation / benchmarking frameworks
• Experience working with large API ecosystems
• Experience with AI observability tools

## Working Style

• Strong ownership and autonomy
• Ability to work from problem architecture implementation
• Comfortable exploring unfamiliar codebases, APIs and product logic
• Focus on engineering reliable AI systems, not just experimenting with models