career

Senior Principal Research Scientist and Science Leader for Foundational Machine Learning

CSIRO

Leading Foundational Machine Learning at CSIRO within the Analytics and Decision Sciences program, setting technical vision and strategy for making modern machine learning a fundamental part of decision-making under uncertainty.

• Co-created Arrow, one of the first foundation models for causal discovery
• Leads research spanning causal discovery and inference, optimal experimental design, and spatio-temporal modelling, underpinned by principled uncertainty quantification
• Translated research into real applications across renewable energy, climate systems, aerospace, and education

Principal Research Scientist

CSIRO

Provided scientific leadership as Team Leader (Foundations and Methods) and Group Leader (Statistical Machine Learning).

• Built AutoGP and VGCN, platforms for scalable Gaussian process modelling and graph-structure learning, turning research advances into reusable infrastructure
• Developed RL-BOED, deep reinforcement learning methods for agentic Bayesian sequential experimental design
• Developed model selection methods for complex Bayesian deep neural networks
• Developed efficient inference algorithms for Gaussian process models and doubly stochastic Poisson processes
• Introduced new approaches to Bayesian optimisation and Bayesian graph convolutional networks

Senior Lecturer

UNSW

Led an independent research program in probabilistic machine learning, focused on scalable Bayesian inference and structured prediction.

• Developed scalable inference methods for nonparametric Bayesian models, including Gaussian processes
• Developed novel techniques in structured prediction and non-conjugate probabilistic inference
• Applied this work to solar energy forecasting, contributing to renewable-energy applications of probabilistic ML

Senior Researcher

NICTA

Carried out research on large-scale inference in Gaussian process models, published at top-tier machine learning venues.

• Developed data fusion methods for the characterisation of geothermal energy targets, in collaboration with the Australian geothermal sector
• Built spatio-temporal models for solar energy forecasting

Researcher

NICTA

Carried out research published at top machine learning venues, and contributed to applied business activities.

• Non-parametric Bayesian models for elicitation of user preferences
• Methods for coherence improvement in statistical topic models
• Activity recognition from wearable sensors
• Core machine learning technology for mining news and blogs
• Delivered a "Practical Machine Learning" course for industry

Research Associate

The University of Edinburgh

Led research and development of compiler technology that can automatically learn how to best optimise programs, and pioneered multi-task Gaussian process methods for transfer learning under uncertainty.

• Developed machine learning techniques able to learn good optimisation strategies across different programs and computer microarchitectures
• Delivered the machine learning model integrated into Milepost GCC, the world's first open-source machine-learning-driven compiler, later recognised with multiple international Test-of-Time awards
• Pioneered multi-task Gaussian process methods for transfer learning under uncertainty, now widely used across machine learning and scientific applications

Project Engineer

Casa Editorial El Tiempo

Developed software applications for managing the cash flow of the company, with technologies such as SAP, Microsoft Visual Basic and SQL Server.

• Developed software applications for managing the cash flow of the company

Software Developer

Inpesa Ltd, Ecopetrol ICP

Investigated and implemented machine learning techniques for the characterisation of oil fields.

• Led the integration and deployment of the software tool "Oil Field Intelligence"

Lecturer

Universidad Industrial de Santander

Teacher for the undergraduate level courses Numerical Analysis and Digital Processing of Speech Signals.

PhD

The University of Edinburgh

Compilers that learn to optimise: A Machine learning approach

MSc with Distinction

The University of Edinburgh

Predicting good compiler transformations using machine learning

B.Sc., Computer Science, Summa Cum Laude

Universidad Industrial de Santander

ICML Oral Presentation

International Conference on Machine Learning (ICML)

Oral presentation (top ~1% of submissions) for "Renyi Neural Processes." Senior author; presented by postdoctoral researcher Xuesong Wang.

Best Research Paper Award (Student)

ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD)

Feature-based Learning for Diverse and Privacy-Preserving Counterfactual Explanations. First authored by student Vy Vo.

Outstanding Contribution

NeurIPS' Graph Representational Learning Workshop

Variational Graph Convolutional Networks

Test-of-time paper award

International Symposium on Code Generation and Optimization (CGO)

Automatic Feature Generation for Machine Learning Based Optimizing Compilation

Test-of-time paper award

International Symposium on Code Generation and Optimization (CGO)

Rapidly Selecting Good Compiler Optimizations using Performance Counters

Faculty of Engineering Silverstar Award

The University of New South Wales (UNSW)

AWS in Education Research Grant

Amazon Web Services (AWS)

Faculty Research Grant Program Award

The University of New South Wales (UNSW)

HiPEAC Paper Award

HiPEAC (European Network on High Performance and Embedded Architecture and Compilation)

European Network of Excellence on High Performance and Embedded Architecture and Compilation award.

MSc in AI with Distinction

The University of Edinburgh

For outstanding performance on the taught courses and dissertation.

Summa Cum Laude

Universidad Industrial de Santander

For having obtained a cumulative GPA superior to that of all students in computer science during the previous five years.