Responsibilities Design and implement automated and manual QA strategies for GenAI systems
Build regression tests, generate synthetic test data, build test harnesses and evaluation pipelines for accuracy, grounding, robustness, safety
Validate prompt logic, model behavior and evaluation metrics
Execute tests and debug pipelines
Support deployment of GenAI products
Lead system design reviews, advocating best practices for architecture and reliability of GenAI applications
Collaborate closely with data scientists and developers to embed testing and quality checkpoints
Establish testing standards and workflows for continuous integration and deployment
Conduct code reviews with a focus on design, testability, and maintainability
Document system design decisions, test cases, and best practices
Close collaboration with domain experts and stakeholders to validate developed solutions
Requirements Strong software engineering background with QA ownership
Familiarity with RAG & Graph frameworks
3+ years of experience as a software engineer (preferably 5),



or QA engineer in production systems, preferably including at least 1 year experience in GenAI systems
Advanced Python skills
Experience working with LLM-based systems in real environments
Hands-on experience with: LLM APIs, AWS Bedrock (Anthropic & OSS), Azure Foundry (OpenAI), GCP (Gemini), Cloud platforms (AWS, Azure, GCP)
Proven ability to implement test frameworks for GenAI applications
Knowledge of API testing, data validation, and data pipeline testing
Experience with version control (Git)
Experience with building and deploying AI/ML products through the full product lifecycle
Good understanding of scientific computing and ML frameworks
Solid understanding of safety, bias, and failure modes of GenAI systems
Comfortable working in a relatively fast-moving environment where priorities may shift and initiative is valued
Proactive problem-solver who can anticipate needs and find solutions
Comfortable working independently,

📌 Senior Engineer (Madrid)
🏢 Jobtailor
📍 Madrid