Predictive AI Data Scientist
Pubblicato il 08-09-2026 - Unicredit in Milano
Join to apply for the Predictive AI Data Scientist role at UniCredit.
We are looking for an experienced Senior Data Scientist to lead the creation of predictive AI solutions, leveraging Machine Learning and Big Data capabilities. You will be involved in the full E2E development lifecycle of our ML solutions, from data architecture design to final delivery and production maintenance. You will also contribute to the development of our internal central assets, such as the feature store and shared libraries, through our MLOps framework. The solutions you will work on will impact important bank processes (e.g. credit evaluation, KYC, CRM) and will be built fully aligned with regulatory requirements (GDPR, AI Act, ECB guidelines).
Responsibilities
- Design and develop predictive AI solutions by applying advanced ML and big data techniques.
- Define data architectures and pipelines to support scalable AI solutions.
- Train, evaluate, and optimize Machine Learning models for real-world applications.
- Develop and maintain centralized AI assets, including feature stores and shared libraries.
- Integrate and manage ML solutions within our MLOps framework for seamless deployment.
- Adopt a squad service model following Spotify schema.
- Embed regulatory requirements fully and stay aware of regulation updates.
- Collaborate with multiple teams across the bank and across countries (Data Platforms, Digital factories, Risk, Finance, Compliance functions, Cross-country CRM) in ideation, development and deployment phase.
- Work in a smart, young team fully engaged in revolutionizing the bank.
Qualifications
- University degree in Computer Science, Engineering, Mathematics, or Statistics.
- 4+ years of relevant experience in Data Science, AI, or ML Engineering.
- Advanced proficiency in Python and experience with big data engineering practices.
- Strong knowledge of ML and deep learning frameworks (TensorFlow, PyTorch, Scikit-learn).
- Hands‑on experience with third‑party data pla