Principal Lead Data Scientist / Applied AI

AI Innovation, within AI Transformation, is responsible for knowing and exploring both new technologies and new Artificial Intelligence models and solutions that may have an impact and adoption by BBVA.

From the creation of the Robots and the definition of The Frame as a key framework and enabler for AI scalability, a great effort is required in terms of both research and innovation, as well as strategic supervision and "advisory" within the organization.

Within AI Innovation, the AI Lab explores, evaluates, and develops advanced solutions based on foundational models, Large Language Models (LLMs), and agentic systems for the bank's strategic use cases. The team works across the entire lifecycle: from experimental model evaluation to the construction of prototypes and their integration into production environments.

We are looking for a person for the position of Principal Data Scientist / Applied AI, with more than 15 years of experience in Machine Learning, GenAI, and in the development of projects and solutions that combine data and analytical and ML models. Advanced analytical experimentation will be combined with the development of solutions based on LLMs and agentic architectures, from use case definition to production deployment.

Team management skills are required to extend existing collaboration with other business areas such as Commercial Banking, CIB, and Finance. International experience and multi-country leadership will also be valued to expand work lines in other countries such as Mexico and Turkey.

• Demonstrated experience in innovation is required
• Select and comparatively evaluate foundational models for different use cases, including language and voice applications and multimodal solutions.
• Design and execute evaluation plans, defining technical and business metrics and comparing models in different scenarios.
• Iteratively build training, validation, and evaluation datasets.
• Adapt generative and discriminative models through fine-tuning or other equivalent techniques.
• Define controlled experimentation methodologies, such as A/B testing, and validate analytical KPIs associated with LLM-based solutions.

Development of LLM-based solutions and agentic systems

• Design and develop analytical solutions based on LLMs, including architectures that combine several models.
• Design and develop agentic systems in which models interact with tools, APIs, external knowledge sources, and ML models.
• Develop reasoning flows and model orchestration, with task planning, tool use, and composition of several inference steps.
• Build knowledge bases and domain datasets to feed generative models.
• Create systems where LLMs function as intelligent interfaces to interact with corporate systems and data.
• Develop inference pipelines and backend services; integrate models with the bank's systems and operations, including APIs and corporate systems.
• Participate in iterative experimentation cycles, from prototype to production solution.
• Continuously explore advances in LLMs, foundational models, advanced NLP techniques, and agentic architectures.
• Evaluate relevant technologies and frameworks for generative AI solutions.
• Communicate technical concepts to internal audiences with different levels of specialization.
• Experience in the full lifecycle of ML and NLP projects: exploratory data analysis (EDA), dataset construction, model selection and training, evaluation, and monitoring.
• Experience developing production analytical solutions, ML pipelines, and inference services.
• Experience in MLOps: pipeline automation, continuous integration and deployment, observability, and model monitoring.
• Experience working in cloud environments, especially AWS and GCP, and deploying and integrating ML solutions in the cloud.
• Knowledge of infrastructure as code (IaC), for example CloudFormation.
• Experience with Git and CI/CD practices for software integration and deployment.
• Programming in Python; experience with TypeScript or Node.js will be valued.
• Experience with LLM orchestration and agent frameworks, such as LangGraph, Agents SDK, AWS Strands, or other equivalents.
• Experience using programming agents such as Codex, Claude Code, Cursor, GitHub Copilot, or similar tools.
• Understanding of risks associated with generative models, including fact hallucination, biases, unpredictability, and omissions in generation.

• Team spirit and ability to work collaboratively.
• Ability to explain technical concepts to non-technical profiles.
• Analytical mindset and orientation toward experimentation. Flexibility and adaptation to evolving technological environments.