Rares Dimitrie Grozavescu

Contact Information

Trinity Street, Trinity College
Cambridge, CB2 1TQ, United Kingdom
E-mail: rg625@cam.ac.uk
Phone: +40 761 343 198

Education

Trinity College, University of Cambridge

PhD in Computational Statistics and Machine Learning
Expected June 2027

Trinity College, University of Cambridge

MEng in Engineering — Distinction (Top 5%)
June 2023

Relevant Coursework

  • Geometric Deep Learning
  • Probabilistic Machine Learning
  • Computational Neuroscience
  • Deep Learning and Structured Data
  • Machine Learning in the Physical World
  • Statistical Signal Processing and Analysis
  • Data Transmission and Inference
  • Mathematical Methods
  • Information Theory and Coding
  • Mathematics of Machine Learning
  • Riemann Surfaces

Experience

Supervisor — University of Cambridge

Ongoing
Delivered supervisions for undergraduate courses including Mathematical Methods, Statistical Signal Analysis, and Bayesian Inference.

Machine Learning Research Intern — Sony, Tokyo

June 2025 – July 2025
Designed an end-to-end music style transfer system using state-of-the-art generative models, including Stable Diffusion and AudioLDM. Developed a web-based application enabling interactive, real-time music transformation.

Machine Learning Research Intern — SilurianAI

March 2025 – May 2025
Implemented state-of-the-art diffusion models for foundational weather forecasting. Developed ensemble forecasting frameworks trained with CRPS-based loss functions.

Junior AI Researcher — DeepSea Technologies

July 2023 – October 2023
Built machine learning models for real-time vessel routing and fuel/time optimisation for both manual and autonomous shipping systems (Cassandra and Pythia platforms).

NLP & Computer Vision Intern — Infosys

June 2022 – August 2022
Prototyped document information extraction pipelines using LayoutLMv3 and DiT, integrating results into a developing Python package.

Data Analysis Intern — Metail

August 2021 – October 2021
Automated preprocessing pipelines for 3D garment data used in neural network training to improve virtual fitting accuracy.

Software Engineering Intern — Eseye

June 2021 – August 2021
Developed an AWS-based billing engine for high-volume IoT data and explored SIM activation techniques without certificate-based authentication.

Projects

Reboot the Earth Hackathon — irriGO

October 2025, Rome
Developed an end-to-end machine learning framework for predicting soil moisture, identifying drought periods, and recommending optimal irrigation schedules for Southern Mediterranean regions.
Used transformer-based architectures integrating weather forecasts, satellite infrared indices, and in-situ sensor data. Designed a mobile application for real-time agricultural decision support.

Physics-Informed Neural Operators for Ice-Sheet Modelling

June 2023
Developed a machine learning framework for Antarctic ice-sheet dynamics by combining multiwavelet neural operators with physical constraints. Designed an inverse modelling approach using Metropolis–Hastings to recover hidden physical parameters relevant for long-term climate predictions.

Awards & Honors

  • Winner, Reboot the Earth Hackathon — October 2025
  • Trinity College Internal Graduate Studentship — October 2023
  • Senior Scholar, Trinity College — October 2021
  • ARM Award, Best Lego Mindstorms Design — October 2019
  • Honorable Mention, International Astronomy & Astrophysics Olympiad — November 2018