∞ λ Δ π

Sophie Wharrie, PhD

I'm Sophie, a postdoctoral researcher working on artificial intelligence (AI) and machine learning (ML) for real-world deployment, based in Sydney, Australia. I combine practical and global experience from 8+ years building production AI/ML systems with technical and creative expertise from foundations in applied mathematics and a PhD in computer science. I'm highly experienced in designing deep learning and probabilistic machine learning algorithms, as well as utilising and adapting foundation models for complex real-world settings. My research interests draw me towards solving AI/ML problems that overcome practical deployment challenges and real-world data complexities.

View Projects Get in Touch

Featured Projects

A Contextual AI System for Personalised AI Agents Supporting Younger Adults With Cancer

This project addresses challenges faced by the rising demographic of early-onset cancer patients (ages 20–49), whose life milestones and long-term health are significantly disrupted by diagnosis and treatment. We developed an iOS mobile health app that integrates consumer wearables (e.g., Apple Watch) with multimodal foundation models to unify fragmented clinical data, signals from physiological sensors, and the reality of daily life with cancer.

The core technical innovation lies in a contextual AI system built on a unifying data model that temporally aligns three categories of data inputs - treatments, patient responses, and additional clinical context. I created algorithms for AI agents to reason more effectively about cause and effect relationships in longitudinal health data. In a 12-month deployment for a 27-year-old patient with triple-negative breast cancer (TNBC), this approach proved effective for the early detection of concerning physiological changes during cancer treatment, tailored summaries and appointment planning, and the management of immunotherapy-related toxicities.

  • Large Reasoning Models (LRMs)
  • Multimodal Foundation Models
  • Generative AI
  • Retrieval-augmented generation (RAG)
  • Causal Inference
  • Multivariate Time Series Algorithms
  • Heterogeneous Time Series Datasets
Read more

PhD Research: Probabilistic Machine Learning and Deep Learning for Health and Genetics

My PhD research was motivated by my practical experience deploying AI systems, recognising the need for general solutions to recurring technical problems. The methods developed in my research are applicable across many domains but I took a particular interest to advancing AI/ML for impact in health, computational biology and personalised medicine, where unique data complexities call for novel approaches.

My research publications present new algorithms for large-scale health and genetics data, validated for genetics biobanks and nationwide health registries in Finland and the United Kingdom: (1) methods and software for generating high-dimensional synthetic data simulations for human genetics research, and (2) advanced deep learning techniques for modelling longitudinal health data. This includes several deep learning algorithms, including a graph neural network based architecture for modelling disease risk in families and a Bayesian meta-learning approach for ensuring that algorithms generalise across diverse environments.

  • Deep Learning
  • Probabilistic Machine Learning
  • Graph Neural Networks
  • Graph (Geometric) Representation Learning
  • Bayesian Meta-Learning
  • Causal Machine Learning
  • Synthetic Data Generation
  • Out-Of-Distribution (OOD) Generalisation
  • Robustness to Distribution Shifts in Deployment
Read more

CV

8+ years of global experience in AI research and engineering:

Experience

  1. Research Fellow — University of Melbourne

    Jan 2025 – Dec 2025 · Melbourne, Australia

  2. Machine Learning Researcher — Finnish Center for Artificial Intelligence

    Dec 2020 – Dec 2024 · Helsinki, Finland

  3. Co-Founder & AI Engineering Lead — Velmio

    Jan 2020 – Dec 2022 · Tallinn, Estonia

  4. Consultant, Data and Artificial Intelligence — Deloitte

    Jan 2018 – Jan 2020 · Sydney, Australia

  5. Earlier Internships: Econometrics Research — Reserve Bank of Australia; Actuarial and Data Science Consulting — PwC

    Pre 2018 · Sydney and Perth, Australia

Education

  1. Aalto University — PhD, Doctor of Science (Technology), Computer Science

    2021 – 2025 · Thesis — Advancing towards personalised medicine: probabilistic machine learning and deep learning for health and genetics · Advisor — Prof. Samuel Kaski

  2. University of Sydney — First Class Honours, Applied Mathematics

    2018 · Awarded the K.E. Bullen Memorial Prize in Applied Mathematics

  3. University of Western Australia — BSc, Mathematics & Statistics, Economics

    2015 – 2017 · High distinction average

Technologies

Tooling I'm currently building with in my projects:

Programming Languages

  • Python
  • C++
  • Julia

ML Tooling and Frameworks

  • PyTorch
  • JAX
  • Weights & Biases
  • MLflow
  • Optuna

Cloud, Data & HPC

  • Slurm
  • Vertex AI (GCP)
  • GCP Cloud Run
  • Docker
  • AWS SageMaker
  • PostgreSQL
  • DuckDB
  • Firebase

Contact