> full description
We’re looking for a Middle/Senior Back-end Developer from Wroclaw, Madrid or Valencia to help build and scale an AI-powered customer support platform. You’ll ramp the way everyone on the team does: your first weeks focus on stability, robustness, and bug-fix work across the live chat deployment — the fastest way to learn the architecture end-to-end while contributing immediate value to production customers. As you come up to speed, you’ll take on feature and framework development with growing scope and ownership. You don’t need to be an AI expert. You need to be a strong, fast-learning engineer who has genuinely built things with LLMs and wants to go deep on production agentic systems.
What You’ll Do
- Build features across the platform: backend services, conversational flows, and components of the agentic harness (tool/function calling, context management, guardrails)
- Extend the tool and adapter layer that exposes the cloud APIs and real-time telemetry to LLMs
- Investigate and fix issues spanning LLM behavior (prompts, tool calls, context) and conventional backend causes (APIs, data, concurrency) — and build the instincts for systems where a “bug” may live in a prompt rather than in code
- Strengthen the platform’s production quality: test coverage, evaluation cases, failure handling, and observability
- Contribute to the platform’s expansion from chat into real-time voice as it develops
What You’ll Bring
- 3+ years of professional software engineering experience building backend systems
- Solid Java experience
- Genuine hands-on experience building with LLMs: you’ve used LLM APIs with tool/function calling, iterated on prompts and context, and can speak concretely about what you built, what broke, and how you fixed it — from production work, internships, or substantial personal projects
- Good fundamentals: data structures, APIs, testing, debugging, and version control; familiarity with cloud services and containers
- Fast, self-directed learning — evidence you can absorb a complex architecture in weeks and then work with increasing autonomy
- Care and rigor appropriate to a production system for Tier-1 enterprise customers handling sensitive subscriber data
- Upper-Intermediate English level
Nice to Have
- Agent frameworks, LLM orchestration libraries, or MCP-style tool integration
- Experience writing evaluations or tests for LLM-driven features
- Event-driven systems (e.g., Kafka), Kubernetes, or CI/CD experience
- Telecom, IoT, networking, or other telemetry-rich domains; interest in voice/real-time AI