Summary / Purpose Join us and build the next generation of travel AI. Power a Chat Assistant used by millions, evolving into an agentic system that goes beyond FAQs to deliver end-to-end solutions. Help create the most intelligent and seamless travel experience in the industry. Key Responsibilities - Craft and fine-tune advanced LLM prompts to drive high-performance conversational experiences in the Chat Assistant - Use a data-driven approach to identify limitations in current conversational models and generate effective solutions for challenges such as user engagement, contextual understanding - Design reusable frameworks - Collaborate closely with engineering teams to align GenAI strategies with backend enhancements, contributing to architectural optimization - Prototype, deploy and maintain new LLM/agentic AI based strategies - Continuously evolve the LHG Chat Assistant architecture with a focus on deepening the assistant’s natural language understanding and response relevance, integrating agentic AI principles to enable more adaptive, goal-oriented interactions in our redesigned system - Build LLM-backed transactional flows - Transition towards agentic AI, orchestrate hybrid solution with AI + deterministic logic - Implement and maintain function calling to backend systems or APIs - Automated testing for all components of AI solutions (prompt, flows, function calls, etc.) Minimum Qualifications - At least 3 years of commercial experience in Data Science or Machine Learning - Expertise in designing, testing, and optimizing solutions for large language models (e.g., ChatGPT, GPT-4) - Proficiency in Python and at least one AI/data- platform (e.g. CognigyAI, databricks) - Hands-on expertise with AI/NLP platforms and performance evaluation (A/B testing, error analysis) - Solid grasp of current AI/ML research, with a focus on best practices for GenAI and emerging agentic AI concepts - Strong collaborator in agile, cross-functional teams, with excellent technical communication - Awareness of ethical AI practices, including bias mitigation and data privacy - A systematic approach to identifying issues in existing setup and constructing targeted solutions, including the ability to troubleshoot and optimize model behavior under different conversational scenarios - Very good command of English - Experience with Scrum teams - Willingness to upskill on your own initiative and deriving further insights from data - Willingness to come to our office in Gdańsk two days a week Preferred Qualifications - Not stated (not separately listed on this posting) Benefits / Compensation (if stated) - Not separately listed in body text captured (Lufthansa Group postings show a generic Benefits icon list on-page not always extractable as text; see source link for full benefits icons) ---