I'm Co-founder of Heartspace, where we're building tools for better breathing. The techique of resonant breathing helps to reduce stress and improve performance under pressure (see our blogs on the subject here). The breathing practise has been studied since the 80s and has recently come into prominence both in academia, and in popular revivals of traditional eastern practises.
Our product teaches deep breathing using a process called Heart Rate Variability Biofeedback using a wearable heart rate monitor and data analytics built into our app. We are currently launching in New Zealand and Australia.
2019 - 2023
I was Director of Data at Halter where I led the data science team, modelling pasture growth and building models for animal position tracking.
This included the data collection process, designing the appropriate model architectures, and reading a lot of papers about cows and grass.
Developing pasture growth models on the Halter Research & Development farm.
In the senior leadership team we made decisions to launch into new markets, initiate new R&D, and cut losses on stagnant projects. We faced demanding growth expectations and tight market conditions while developing a challenging product that was mission-critial for our customers. We set up relationships to run studies on virtual fencing with world leaders in the field.
Kicking off the collaboration with Tasmania Institute of Agriculture for research on dairy farm productivity.
Early on I developed the strategy for critical systems projects including how to calibrate a core sensor efficiently across on thousands of devices and conserve energy across the fleet, enabling Halter to deploy into new markets.
I joined Halter after the Series A round of funding, and co-ordinated the first farm deployment. Halter has since deployed on hundreds of farms, in NZ and internationally and raised Series B & C funding rounds (USD $74M total).
2015 - 2019
My PhD in biomedical engineering explored how needle free drug delivery could be controlled more accurately using laser imaging. I showed the depth of the injection could be calculated by measuring the scattered light pattern produced from light in the water jet. I built a laser-jet injection device, simulated the optics of skin tissue, and solved the inverse problem to calculate injection depth from scattered laser-light patterns.
My colleagues and I with the jet injector system, which delivers vaccines in a fraction of a second without a needle.
In 2016 I was a visiting researcher at Massachusetts Institute of Technology, on a 3-month exchange. During my PhD I published 3 journal papers, presented at international conferences and my work was nominated for thesis of the year.
A presentation on my thesis for the UoA 3-minute thesis competition.
My PhD brought my background in mathematical modelling together with my passion for applied engineering in the biomedical domain. I developed a greater appreciation for hardware and electronics and honed my skills in experimentation and iterative development.
2014 - 2018
As a Research Engineer at Soul Machines under Dr. Mark Sagar, I developed a neural network for motor control of a virtual human avatar. The avatar could be trained to draw and play pong, built on a simple network called a self-organising map.
Demonstrating the interactive avatar, BabyX, which is animated by neural networks.
I learnt to build models in a streaming, 3D graphics environment, and continued to work for Soul Machines during my PhD. When I began at Soul Machines in 2014 we were a team of six. By the time I left the company was over 70 employees and had raised USD $7.5m. As of 2024, Soul Machines has raised USD$135m in funding.
2012 - 2014
BE at The University of Auckland, top of class in Engineering Science. This is where I developed a passion for applied mathematical modelling. My honours project was in the research lab which would later become Soul Machines. Optimisation problems were a theme throughout the degree, a notable project involving the simulation of power generation of a wind turbine, which we validated by building and testing it at the university's wind tunnel.
Testing the wind turbine design at the wind tunnel
Community involvement has played an important role in shaping how I like to lead teams. Through my undergraduate degree I tutored Physics and Biology in low socio-economic high schools, and was a supervisor in a hostel for first-year and international students.
Finishing up the year with the students at International House, UoA and the resident advisor team.
It's possible to control your stress response through controlled breathing, with a technique called Biofeedback Training. I built Every-Breath-You-Take to measure and train this with a Polar H10 heart rate monitor. The project received some attention after being on the front page of Hacker News for a week and formed the basis for building Heartspace.
Atrial Fibrillation (AF) is a heart beat arrhythmia that affects 30 million people worldwide. I built To Beat or not to Beat to detect AF with a wearable heart rate monitor. The project is an implementation of recent paper on the topic and uses a convolutional neural network trained on a public electrocardiogram dataset.
Heart rate awareness – the ability to perceive your own heartbeat – correlates highly with emotional control and pain control. I built Interoception to measure and train this with a wearable heart rate monitor. It's an implementation of a widely used protocol in the field of Interoceptive Awareness.
In 2018 I started The Rest Project with a friend to help people use flotation therapy to reduce stress. Hundreds of people around NZ used our programs to manage stress.
