- Ioannis Zachos(aka Yannis)Machine Learning Scientist (PhD)
Machine Learning Scientist with a Cambridge PhD and 6 years of experience engineering AI solutions to complex problems. Expertise in solving high-dimensional data challenges using Python and Generative AI, evidenced by State-of-the-Art contributions at NeurIPS '24. Passionate about bridging the gap between R&D and production.
Professional Experience
Cambridge University & Arup City Modelling Lab
Machine Learning Reasearcher10/2020 - 10/2025, Cambridge, UK
- Leveraged a collaborative industrial partnership to build simulation frameworks for urban planning.
- Engineered a State-of-the-Art framework for multi-agent origin-destination trip modelling leveraging Neural Stochastic Differential Equations.
- Designed and implemented a custom Markov Chain Monte Carlo sampler to solve high-dimensional inverse problems in multi-agent system simulations.
- Developed an open-source tool for synthesising origin-destination trips, enabling Arup engineers to simulate realistic transport scenarios.
- Leveraged a collaborative industrial partnership to build simulation frameworks for urban planning.
- Engineered a State-of-the-Art framework for multi-agent origin-destination trip modelling leveraging Neural Stochastic Differential Equations.
- Designed and implemented a custom Markov Chain Monte Carlo sampler to solve high-dimensional inverse problems in multi-agent system simulations.
- Developed an open-source tool for synthesising origin-destination trips, enabling Arup engineers to simulate realistic transport scenarios.
Cervest Ltd (now Mitiga Solutions)
Statistical Scientist09/2018 - 07/2019, London, UK
- Key contributor to the 0-to-1 engineering lifecycle of an AI-powered climate intelligence platform.
- Architected and developed scalable ETL pipelines in Python to ingest, process and fuse ~10TB of multi-resolution multi-modal satellite imagery.
- Engineered and deployed spatio-temporal ML models for environmental resilience.
- Translated complex modeling results and predictions into actionable insights for investors and clients, facilitating technical due diligence.
- Key contributor to the 0-to-1 engineering lifecycle of an AI-powered climate intelligence platform.
- Architected and developed scalable ETL pipelines in Python to ingest, process and fuse ~10TB of multi-resolution multi-modal satellite imagery.
- Engineered and deployed spatio-temporal ML models for environmental resilience.
- Translated complex modeling results and predictions into actionable insights for investors and clients, facilitating technical due diligence.
Eurobank Private Bank Luxembourg
Investment Advisory Intern06/2018 - 08/2018, Athens, GR
- Designed and deployed an R Shiny application that performs on-demand portfolio optimisation subject to diversification and volatility constraints.
- Designed and deployed an R Shiny application that performs on-demand portfolio optimisation subject to diversification and volatility constraints.
iQom Ltd (now Epsilon Net)
Data Analyst Intern08/2016 - 09/2018, Thessaloniki, GR
- Analysed high-volume CRM data using R to model customer call arrival times using Poisson processes, optimizing support center resource allocation.
- Analysed high-volume CRM data using R to model customer call arrival times using Poisson processes, optimizing support center resource allocation.
Education
University of Cambridge
PhD in Computational Statistics and Machine Learning
11/2020 - 10/2025
- Industrial collaboration with Arup City Modelling Lab.
- Lead author: NeurIPS 2024 and Stat (Wiley).
- Award: Full scholarship by Arup and EPSRC.
- Thesis: Inverse Problems in Agent-Based Models of Spatial Phenomena.
- Supervisors: Prof. Mark Girolami and Prof. Theodoros Damoulas.
University of Cambridge
MRes in Engineering (Distinction)
10/2019 - 11/2020
- Focus: Computational Statistics and Machine Learning.
- Award: Full scholarship by Arup and EPSRC.
University of Warwick
BSc in Data Science (1st class)
10/2015 - 07/2018
- Thesis: Bayesian online change-point detection for time series segmentation and forecasting in point processes (79%).
- Courses: Machine Learning (73%), Mathematical Statistics (79%), Topics in Data Science (82%), Artificial Intelligence (72%).