Overview
Marcos is working on two strands: creating Claude Projects to guide students on his Databases course through database modelling and implementation. His second strand focuses on how agentic AI frameworks can be used for teaching in quantitative courses, especially those involving the design of data science and machine learning programs.
Key work so far
Using Claude Projects
During Autumn Term 2025, Marcos created Claude Projects to guide his Databases students through database modelling and implementation. AI agents act as critical reviewers of database models and help students produce the necessary code to create and manipulate the database. Feedback was collected via a survey and full implementation of the redesigned course is planned for the 2026/27 academic year.
Agentic AI frameworks for data science
Marcos is exploring how agentic AI can be taught and used in quantitative subjects, so students learn to supervise autonomous agents for data engineering and code-generation tasks – reflecting emerging industry practice.
His aim is to create practical teaching examples that demonstrate how students can work alongside agents to shift their learning focus from repetitive coding "typing" to problem understanding and orchestration and coordination tasks. The agents will help students to understand a data science problem, generate implementation plans and related code, test and validate solutions, and generate documentation and explanations.
His aim is to produce practical, reusable teaching examples built on open-source or free-to-use tools, so approaches can be adopted more widely by other educators.
Impact
Feedback from the Autumn Term pilot was positive, showing that AI agents really help students to concentrate on understanding the problem and critically think about the solution instead of spending time in repetitive, mechanical code tasks.
What’s next?
The new teaching materials and examples will be rolled out from 2026/27 in selected courses in the Department of Statistics. Examples include:
-
database modelling and implementation targeting different types of databases, such as e-commerce, financial and stock market, and healthcare settings
-
data engineering pipelines, where students use agents to help in data acquisition and preparation task
-
machine learning pipelines for classification and regression tasks – as these are fundamental skills in data science