cv

General Information

Full Name Natalia Ślusarz
Languages English, Polish

Education

  • 2021 - 2026
    PhD
    Heriot-Watt University, Edinburgh, UK
    • My thesis focused on the intersection of ML and formalisation ‑ I investigated a method of property‑based NN training, differentiable logics, by applying a more rigorous approach and theoremprovers (Rocq and MathComp)
    • I have workedon the development of this training method ina Haskell‑based NN verification tool ‑ Vehicle.
  • 2017 - 2021
    BsC (Hons) in Mathematics and Computer Science
    Heriot-Watt University, Edinburgh, UK
    • Finished with First Class Honours.
    • Topic.
      • My thesis focused on mathematical properties of neural networks, investigating the influence of complexity of data on the behaviour of the learnt NN representation.

Experience

  • 2026 - CURRENT
    Research Associate (CRADLE)
    University of Manchester, UK
  • XII 24 - IV 25
    Research Asisstant
    Heriot-Watt University, Edinburgh, UK
    • Worked as part of a project AISEC (AI Secure and Explainable by Construction).
    • My work was closely tied with my PhD and focused on providing a formalisation of Differentiable Logics, a neural network training method being implemented in the verification tool Vehicle (which has been developed as part of the AISEC project).
  • 2021 - 23
    Teaching Assistant
    Heriot-Watt University, Edinburgh, UK
    • Delivered labs in multiple subjects ranging from ML, through programming languages to academic writing - Data Mining and Machine Learning, Advanced Software Engineering, Software Development, and Research Methods and Project Planning.
    • Assessed students' coursework and lab projects, advised on Master's projects.
  • X 21 - III 22
    Research Associate
    Heriot-Watt University, Edinburgh, UK
    • Worked as a part-time RA on a research project ``Neural Network Verification; in search of the missing spec''
    • The goal of this project was to address the current under-defined semantics of neural networks. My responsibilities focused around identifying methods from other research areas that could be applied for the purposes of neural networks.
  • VI 2021 - VIII 2021
    Intern in Technology R&D
    PwC Poland | Financial Crime Unit, Gdynia, Poland
    • Worked on tool development in the Research and Development Technology team and improved the tool's performance by using neural networks to meet industry standards
    • Contributed towards implementing modern machine learning solutions in new internal tools

Other Academic Activities

  • 2026
    Publicity Chair of ITP'26
    • In charge of communication with conference attendees and PC members, as well as maintaining the conference website
  • 2025
    Program Committee Member for Rocqshop
    • Reviewed submissions for the workshop (previously known as Coq Workshop)
  • 2024-2025
    Co-organiser of Quantitative Logic Reading Group
    • Co-organised a series of reading group seminars related to quantitative logic, held in a blended format
    • Was in charge of leading the discussions and organisational matters
  • 2020-2025
    Member and Volounteer | LAIV (Lab for AI and Verification)
    • Gained expertise in AI verification, especially related to neural networks. Had an opportunity to present and discuss my work with other experts in my field
    • Provided help with organisation of a seminar series by taking over the seminar schedule and notifications, as well as management of the seminar-wide mailing list
  • IV 2025
    Volounteer | Theorem Proving and Machine Learning in the age of LLMs; SoA and Future Perspectives
    • Coordinated the online stream of the workshop as well as handled the technical setup and testing
  • X 2023
    Volounteer | Workshop on Safe and Robust Machine Learning
    • Helped with organisation of the local venue and technical setup
  • VII 2022
    Volounteer | Scottish Programming Languages and Verification Summer School
    • Handled the technical setup of the lectures, especially the coordination of the live stream delivered in a blended format