AI Automation Engineer
On-siteŁódź, Łódź Voivodeship, Poland
Job Summary
Build multi-step agent systems that operate across business applications, creating typed tools and API wrappers to connect AI agents to production systems via the Model Context Protocol. Write comprehensive unit tests and design evaluation pipelines using LLM-as-a-judge frameworks to ensure agent accuracy, safety, and reliability. Monitor agent performance and track regressions using observability tools while implementing sandboxing and defenses against prompt injection. Plan and introduce intelligent AI Agents that enhance and automate operations across QA, Project Management, and Sales within the first six months.
Required Qualifications
- Hands-on experience or familiarity with multi-agent and orchestration libraries such as LangGraph, LangChain, CrewAI, or AutoAgents / AutoGen
- Familiarity with integrating LLMs (e.g., Gemini, Claude) and leveraging cloud-native AI builders like AWS Bedrock or Azure AI Foundry
- Understanding of prompt engineering principles and LLM failure modes (hallucination, drift)
- Basic knowledge of Git version control, CI/CD pipelines, containerization (Docker, Kubernetes), and Agile methodologies
- Understanding of AI safety best practices, including sandboxing, defenses against prompt injection, and secure credential handling
- Ability to work effectively in ambiguous and rapidly changing environments
- Strong analytical thinking and curiosity about emerging AI technologies
- Ability to evaluate non-deterministic AI outputs and make informed technical decisions
- Ability to communicate technical concepts and trade-offs to both technical and non-technical stakeholders
Desired Qualifications
- If you have a portfolio or GitHub profile showcasing your work, we would love to see it. Feel free to share examples of developed agents, along with any details about the architecture, tool-calling logic, and how you evaluated performance.
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.