Machine Learning and Healthcare Transformation

I am a Senior AI & Data Scientist in Singapore’s Ministry of Health Office for Healthcare Transformation, working on system-level healthcare innovation. I am a machine learning scientist and engineer by training, and my current work contributes to several health-tech and population health initiatives. I am also active in the basic machine learning research community.

From a machine learning perspective: My work usually favors methods from Bayesian inference and deep learning. Recent general themes include the modeling of highly-structured time series and relational networks, analyzing clinical trial and electronic health record data, and modeling the behavior of individuals with data from multiple sensors.

From a healthcare perspective: We run population health programs, partner on clinical trials, and develop digital health platforms, among other initiatives. My current efforts revolve around mental health, chronic diseases, and the deployment of digital health solutions like digital phenotyping.

Working with our team

Our Data, Science, & Technology team seeks interns, graduate students, and full-time hires across a range of needs. Those that would work directly with me include data scientists as well as those focused on the design and evaluation of population health programs that use innovative technologies. Please see below for more details.


I completed my PhD in the Machine Learning Group at the University of Cambridge, under the supervision of Zoubin Ghahramani. My work around that time focused on Bayesian nonparametrics, network modeling, Bayesian deep learning, MCMC, and variational inference. During my PhD, I did a few internships and consulted for companies, and more recently I was a Vice President & Executive Director at Goldman Sachs in Hong Kong, where I helped to develop a machine learning team in the Securities division.


Work or collaborate with us

We are looking for interns and graduate students to work with us. Those that would be working with me include data scientists and machine learning researchers. Internships can last from three months to one year, and can start at any point throughout the year. I am particularly keen to work long-term with PhD students in machine learning with methodological interests similar to mine. Such collaborations could take the form of repeated summer internships and/or collaborations throughout your degree. Masters students and final year undergraduates with strong training in data science would also be welcome for internships. We also find regular need for students from a public health background (particularly those working in the design and analysis of population health programs, digital health, mental health, and health economics), though interest and at least some training in data analysis, as well as coding skills (preferably in Python), would likely be necessary.

If interested, please feel free to email me with your CV.

We have close, vibrant partnerships with several institutions across Singapore and beyond including the Institute of Mental Health, the Health Promotion Board, clinicians and researchers across health clusters and Universities, tech companies, and start-ups.


Association between wrist wearable digital markers and clinical status in Schizophrenia. 
W. Martanto, Y. Y. Koh, Z. Yang, C. Heaukulani, X. Wang, N. A. A. Rashid, A. Sim, S. Zheng, C. Tang, S. Verma, R. J. T. Morris, J. Lee 
General Hospital Psychiatry (Letter to the Editor), Vol. 70, 134–136, 2021.

HOPES – An integrative digital phenotyping platform for data collection, monitoring and machine learning. 
X. Wang, N. Vouk, C. Heaukulani, T. Bhuddika, W. Martanto, J. Lee, and R. J. T. Morris 
Journal of Medical Internet Research, Vol. 23, No. 3, 2021.

Scalable Bayesian dynamic covariance modeling with variational Wishart and inverse Wishart processes. 
C. Heaukulani and Mark van der Wilk 
NeurIPS, 2019.
[paper] [code]

Variational inference for neural network matrix factorization and its application to stochastic blockmodeling. 
Onno Kampman and C. Heaukulani 
ICML 2019 Workshop on Learning and Reasoning with Graph-Structured Data. 
[paper] [code]

Black-box constructions for exchangeable sequences of random multisets. 
C. Heaukulani and Daniel M. Roy 
ArXiv:1908.06349, Aug 2019.

Bayesian inference on random simple graphs with power law degree distributions. 
Juho Lee, C. Heaukulani, Zoubin GhahramaniLancelot F. James, and Seungjin Choi 
ICML, 2017.

Generalized IBPs, random multisets, and tree-structured feature allocations. 
PhD Thesis. 
University of Cambridge, Jan 2016.

Gibbs-type Indian buffet processes. 
C. Heaukulani and Daniel M. Roy 
Bayesian Analysis, Vol. 15, No. 3, 683–710, 2020. (First appeared Dec 2015.)

Beta diffusion trees and hierarchical feature allocations. 
C. Heaukulani, David A. Knowles, and Zoubin Ghahramani 
ICML, 2014. Extended version.

The combinatorial structure of beta negative binomial processes. 
C. Heaukulani and Daniel M. Roy 
Bernoulli, Vol. 22, No. 4, 2301–2324, 2016. (First appeared Dec 2013.)

Dynamic probabilistic models for latent feature propagation in social networks. 
C. Heaukulani and Zoubin Ghahramani 
ICML, 2013.

Talks & Events

I co-organized the ICML 2020 workshop on Healthcare Systems, Population Health, and the Role of Health-tech, with Konstantina PallaNiranjani PrasadKatherine Heller, and Marzyeh Ghassemi
Virtual, July 2020.
[website][recorded talks]

Some Bayesian extensions of neural network-based graphon approximations. 
EcoSta 2018, Hong Kong, June 2018 
Microsoft Research Cambridge, Cambridge, UK, August 2018 
Uber AI, San Francisco, June 2019

Random partition based inference schemes for feature allocations. 
BNP 10, Raleigh–Durham, June 2015

Beta diffusion trees. 
ICML 2014, Beijing, June 2014

Probabilistic latent feature propagation in social networks. 
ICML 2013, Atlanta, June 2013 
NetSci 2013 Satellite Symposium, Copenhagen, June 2013

The negative binomial IBP. 
BNP 9, Amsterdam, March 2013