R4R (Reproducibility for R) aims at making R notebooks reproducible by detecting sources of non-determinism in the notebook, and creating artifacts from the detected dependencies.
This project has received funding from the European Union’s Horizon Europe research and innovation program, ERC PoC 2022, under grant agreement No. 101081989.
The RiGiD project lays the groundwork for this research programme and aims to develop a methodology for rigorous engineering of data analysis pipelines that can be adopted in practice. Our approach is pragmatic. Rather than chasing functional correctness, we hope to substantially reduce the incidence of errors in the wild. The research is structured in three overlapping chapters: a catalog of error patterns as well as a labeled dataset to be shared with other researchers, a methodology and tooling for developing data sciences codes with reduced error rates, evaluatation by conducting user studies and developing tools for automating deployment.
This project is supported by the Czech Science Foundation under grand program GX23-07580X (excellence in research EXPRO).
The Evolving Language Ecosystems project explores the fundamental techniques and algorithms for evolving programming languages and their ecosystems. Our purpose is to reduce the cost of wide-ranging language changes and obviate the need for devising entirely new languages. Our findings will grant both researchers and practitioners a greater degree of freedom when experimenting with new ideas on how to express computation.
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 695412.
More information at https://ele-prl-prg.github.io
Computer code is increasingly a shared resource. Web sites such as GitHub and BitBucket host tens of millions of software projects. With great amounts of code, come great opportunities and challenges.The Big Code project aims to automatically extract insights from large code bases by a combination of static program analysis and machine learning. The purpose of the project is to address three challenge problems: language ecosystem evolution, predictive workload performance modelling, synthesis of personalized programming hints.
This project is supported by the Czech Ministry of Education, Youth and Sports from the Czech Operational Programme Research, Development, and Education, under grant agreement No. CZ.02.1.01/0.0/0.0/15_003/0000421.
More information at https://bigcode-prl-prg.github.io
In 2026 the lab was awarded three one-year projects in the Internal Competition for Development Projects on the Integration of AI Tools into Research, Creative, and Educational Activities of the Faculty of Information Technology, CTU in Prague. All three study where large language models genuinely help in our work, and where they do not.
TypeWise investigates empirically whether optional type annotations and type checking help large language models generate better code. We measure not only correctness, but also how many repair iterations a type-checker feedback loop saves, how reproducible outputs are across repeated generations, and whether the extra prompt tokens and checking steps pay for themselves. The benchmark covers Python, R with the optional type checker developed in the lab, and Racket, a language that is far less dominant in LLM coding benchmarks.
Principal investigator: Pierre Donat-Bouillud.
Courses in the Programming Languages specialization use assignments such as interpreters, program analyses and code generators, which conventional autograders can only partially assess: tests check behaviour, but say nothing about design, testing strategy, or conceptual misunderstandings. AIPLS develops and evaluates human-in-the-loop AI assistance for continuous feedback on such assignments, piloted in NI-APT, NI-APR and NI-GEN. Final assessment stays with the teacher; the aim is more frequent and timelier feedback, not autonomous grading.
Principal investigator: Filip Křikava, with Pierre Donat-Bouillud and Petr Maj.
GrantPilot examines how far LLMs can take over the administrative side of grant writing, such as turning call documents into checklists, checking structure and budget consistency, and producing pre-submission review reports, without touching the research idea itself. The project compares frontier cloud models with locally deployed ones, documents what each targeted call permits in terms of AI use, and delivers a reusable starter kit for FIT researchers.
Principal investigator: Pierre Donat-Bouillud.