XTB is a global company from the financial industry, focusing on online trading of financial instruments. They are the largest FinTech in Poland and a leader in Central and Eastern Europe. As a Senior AI Engineer, you will be the technical authority for their LLMOps and Agentic strategy, building resilient, enterprise-grade systems where evaluation-driven development and full-stack observability are core requirements.

Responsibilities

Design and scale resilient, end-to-end architectures for LLM and agentic solutions within a high-load, low-latency fintech environment
Act as a strategic technical partner to the Head of AI, leading architectural reviews and Buy vs Build analyses for greenfield AI projects
Take full ownership of the AI product lifecycle, establishing rigorous standards for LLMOps, FinOps, and automated deployment pipelines from PoC to production
Spearhead the adoption of Agentic Evals and RAG-eval frameworks to turn fine-grained observability signals into actionable system improvements
Collaborate with Data and Product teams to architect AI-ready data layers and MCP servers, ensuring seamless agentic access to semantic layers in Snowflake and Databricks
Foster a culture of technical excellence by mentoring engineers and defining best practices for AI-assisted development (Claude Code, Cursor) to elevate team productivity
Bridge technical complexity with business ROI, ensuring all solutions adhere to financial AI Governance and security standards

Requirements

5+ years of experience in architecting and deploying production-grade AI/ML solutions in Azure, GCP, or AWS ecosystems, with mastery of Python and full ML lifecycle
Deep hands-on expertise with LLM orchestration and data frameworks like LlamaIndex, LangGraph, or PydanticAI and building complex, multi-step agentic workflows
Practical experience in developing and deploying Model Context Protocol (MCP) servers to expose enterprise tools and internal data to AI agents
Strong background in MLOps and LLMOps pipelines, including model versioning, automated deployment, and full-stack observability for deterministic and agentic services
Practical experience in implementing evaluation-driven development using frameworks like RAG-eval and agentic evaluations to measure reasoning and tool-calling accuracy
Proficiency in building high-performance APIs with FastAPI and designing semantic layers on top of enterprise data platforms such as Snowflake or Databricks, integrated with vector databases (e.g., Weaviate, pgvector)
Ability to perform Buy vs Build analyses and translate business requirements into technical architectures that balance innovation with ROI and FinOps
Experience in leading technical initiatives, conducting rigorous code reviews, and mentoring other engineers
Excellent communication skills and ability to work cross-functionally with data engineers and product managers in English

Nice to Have

Mastery of Claude Code, Cursor, or GitHub Copilot and a track record of implementing AI-assisted engineering best practices
Previous experience in fintech, specifically within trading platforms, risk modeling, or high-frequency data environments
Familiarity with Small Language Models (SLM) and techniques for strategic cost and latency optimization
Solid understanding of Data Engineering fundamentals, including Spark, Kafka, and modern Lakehouse architectures

Benefits

Real impact on the development of the company and the product
Work in an experienced team that shares knowledge
Regular feedback and clear career paths
Training budget for courses and conferences
Extra day off on birthday and for parents
Equipment tailored to needs
Private medical care and group insurance
Access to e-learning platform and benefits platform
Access to wellbeing platform with workshops and private therapy sessions
Remote work, from office in Warsaw, or coworking space
Regular team-building meetings

Salary: 19,000-22,500 PLN gross/month.

Location: Wroclaw, Poland (hybrid, with remote/Warsaw/coworking options).

Databricks Agentic AI Engineering Jobs in EU

Get jobs like these weekly