NEWSROOM
AI venture MACSO showcases how research-informed start-ups thrive
After spending ten years fully immersed in academia – culminating in a PhD in Applied Mathematics from Waipapa Taumata Rau, the University of Auckland, no less – it would be understandable if machine learning engineer Hwan Goh had found the transition from study to start-up a challenging one. However, “it was actually quite fun,” recalls Hwan, who’s now head of machine learning at pioneering AI company MACSO. “I thought it would be a lot scarier and a lot more difficult.”
Hwan credits a commercialisation course for PhD students, delivered by the Business School’s Centre for Innovation and Entrepreneurship (CIE), with easing his transition from an academic environment to a commercial one. “I took the course in 2016 and I guess it just switched on that idea in my brain: you’re making something scientific – how do you get it out in the world? It’s a massive question and it was always on my mind while I was doing my PhD. When I went into this commercial world, it answered all those questions. That’s why it felt really natural – I was asking those questions anyway, so I was prepared. And a lot of that is attributed to that commercialisation course.” He’s especially grateful to his PhD supervisor, world-leading applied mathematician and entrepreneur Professor Jari Kaipio, for encouraging him to take the course. “He went out of his way to see what resources the University offered outside the mathematics department. I owe almost every success in my career to him, and aspire to be half the scientist that he is.”
Hwan joined MACSO in July 2021, only a few months after the company’s launch. Founded by computer scientist and AI pioneer Saba Samiei, the start-up’s point of difference lies in their focus on sensory AI, says Hwan. “A lot of people focus on human-based artificial intelligence with very large models, for example ChatGPT,” he explains. “What people neglect is how artificial intelligence can be applied to environmental data. So instead of human pictures or speech, we work with things like aerosols, sounds, touch – physical properties. These can be performed with tiny neural networks, so they don’t require massive, expensive servers. That’s what we focus on at MACSO – developing neural networks that help listen in on the environment, and provide interpretations of environmental data using artificial intelligence.”
MACSO are thriving as a provider of data-driven innovations, with solutions developed to address a wide range of issues related to agtech, data collection, air quality and hardware. They recently successfully closed on a second round of funding that was oversubscribed at $2.4 million.
Hwan was the third employee to join MACSO, and built the start-up’s technical platform from the ground up. “We didn’t have a single line of code to start with,” he recalls. “I had to figure out what we should make and then build it from scratch – from the infrastructure to the actual product.” The team has now expanded to six, three of whom are engineers like Hwan. Hwan manages the engineers and is also part of the leadership team (alongside Saba and sales and business manager John Griffith), providing “a technical voice” on opportunities. “Once Saba and John assess the financials behind an opportunity – what is the addressable market, is this within our field – they bring me in on the technical side,” he explains. “I assess whether we can actually do this – would we have to go out of our way, or can we leverage what we’ve already developed?” It’s an aspect of the role that he particularly enjoys. “The back-and-forth – it’s pretty fun. We all put our skill sets and experience into the pool to come up with solutions.”
A sound sensory solution that monitors herd health in pig farms has been a highly successful innovation for the company. The ingenious device, which listens for sounds of respiratory illness in the animals, is currently in use on a large pig farm in Spain, as well as on farms in New Zealand and America. There are plans for deployment in China, too. “We’re across a lot of continents,” enthuses Hwan. “We’ve also connected with one of the largest animal pharmaceutical companies in Europe, who are currently conducting a trial in Spain.”
MACSO’s latest innovation is an AI tool that counters vaping in schools, developed in partnership with US-based tech company Piera Systems, which manufactures aerosol sensors. The partnership was sparked by a conversation between one of MACSO’s board members and Piera’s CEO about the capabilities of Piera’s sensors: while the sensors can detect particles, they aren’t able to interpret what signature the particles represent. “We then realised vaping is a growing problem in almost every single country,” recalls Hwan. “And schools in particular are struggling with the problem, because they don’t have a way to detect vaping. You can’t put cameras in school bathrooms, and students can use deodorants to mask the smell of a vape.” Quick to identify an opportunity to make a positive impact on adolescent health, MACSO developed a model that can differentiate between deodorant, smoke and vape, and has applied this to Piera’s aerosol sensors. Early trials have yielded excellent results.
Although Hwan claims to be “lackadaisical” about his career path, it’s clear that a passion for commercialisation – ignited early by that influential course at CIE – drives him. He emphatically rejects any assumption that the worlds of academia and entrepreneurship are mutually exclusive, citing an experience in Seattle that crystallised his thinking: “I asked some brilliant academics about the commercialisation aspect of the algorithms they were discussing, and they all looked at each other and laughed. That really bothered me – I thought, a lot of taxpayer money and resources are supporting you in making these algorithms, so why are you laughing about the idea of commercialisation? Commercialising your product is almost like a service to society.”
He’s equally passionate about the importance of keeping an open mind in terms of academic or career choices. “Don’t box yourself in,” he advises other researchers and doctoral students. “Don’t say ‘I’m a mathematician, so I do maths’, or ‘We’re academics, so we don’t think about commercialisation’. You’re all part of the same world, and to put up these false boundaries is just unproductive for everyone.”
For Hwan, one of the most rewarding aspects of his career journey has been realising the real-world relevance of his years of intense study. “I finished my PhD when I was 28,” he notes. “That’s a good chunk of your time learning. And it feels great to realise that that learning can go beyond academia, and watch it come into action. Algorithms I wrote, using knowledge I gained from my PhD, are now sitting on a farm in Spain. I just don’t know what can be more fulfilling than that.”
Applications now open for CIE’s Research to Innovation Hatchery – free for all researchers at the University of Auckland.
Hwan Goh
After spending ten years fully immersed in academia – culminating in a PhD in Applied Mathematics from Waipapa Taumata Rau, the University of Auckland, no less – it would be understandable if machine learning engineer Hwan Goh had found the transition from study to start-up a challenging one. However, “it was actually quite fun,” recalls Hwan, who’s now head of machine learning at pioneering AI company MACSO. “I thought it would be a lot scarier and a lot more difficult.”
Hwan credits a commercialisation course for PhD students, delivered by the Business School’s Centre for Innovation and Entrepreneurship (CIE), with easing his transition from an academic environment to a commercial one. “I took the course in 2016 and I guess it just switched on that idea in my brain: you’re making something scientific – how do you get it out in the world? It’s a massive question and it was always on my mind while I was doing my PhD. When I went into this commercial world, it answered all those questions. That’s why it felt really natural – I was asking those questions anyway, so I was prepared. And a lot of that is attributed to that commercialisation course.” He’s especially grateful to his PhD supervisor, world-leading applied mathematician and entrepreneur Professor Jari Kaipio, for encouraging him to take the course. “He went out of his way to see what resources the University offered outside the mathematics department. I owe almost every success in my career to him, and aspire to be half the scientist that he is.”
Hwan joined MACSO in July 2021, only a few months after the company’s launch. Founded by computer scientist and AI pioneer Saba Samiei, the start-up’s point of difference lies in their focus on sensory AI, says Hwan. “A lot of people focus on human-based artificial intelligence with very large models, for example ChatGPT,” he explains. “What people neglect is how artificial intelligence can be applied to environmental data. So instead of human pictures or speech, we work with things like aerosols, sounds, touch – physical properties. These can be performed with tiny neural networks, so they don’t require massive, expensive servers. That’s what we focus on at MACSO – developing neural networks that help listen in on the environment, and provide interpretations of environmental data using artificial intelligence.”
MACSO are thriving as a provider of data-driven innovations, with solutions developed to address a wide range of issues related to agtech, data collection, air quality and hardware. They recently successfully closed on a second round of funding that was oversubscribed at $2.4 million.
Hwan was the third employee to join MACSO, and built the start-up’s technical platform from the ground up. “We didn’t have a single line of code to start with,” he recalls. “I had to figure out what we should make and then build it from scratch – from the infrastructure to the actual product.” The team has now expanded to six, three of whom are engineers like Hwan. Hwan manages the engineers and is also part of the leadership team (alongside Saba and sales and business manager John Griffith), providing “a technical voice” on opportunities. “Once Saba and John assess the financials behind an opportunity – what is the addressable market, is this within our field – they bring me in on the technical side,” he explains. “I assess whether we can actually do this – would we have to go out of our way, or can we leverage what we’ve already developed?” It’s an aspect of the role that he particularly enjoys. “The back-and-forth – it’s pretty fun. We all put our skill sets and experience into the pool to come up with solutions.”
A sound sensory solution that monitors herd health in pig farms has been a highly successful innovation for the company. The ingenious device, which listens for sounds of respiratory illness in the animals, is currently in use on a large pig farm in Spain, as well as on farms in New Zealand and America. There are plans for deployment in China, too. “We’re across a lot of continents,” enthuses Hwan. “We’ve also connected with one of the largest animal pharmaceutical companies in Europe, who are currently conducting a trial in Spain.”
MACSO’s latest innovation is an AI tool that counters vaping in schools, developed in partnership with US-based tech company Piera Systems, which manufactures aerosol sensors. The partnership was sparked by a conversation between one of MACSO’s board members and Piera’s CEO about the capabilities of Piera’s sensors: while the sensors can detect particles, they aren’t able to interpret what signature the particles represent. “We then realised vaping is a growing problem in almost every single country,” recalls Hwan. “And schools in particular are struggling with the problem, because they don’t have a way to detect vaping. You can’t put cameras in school bathrooms, and students can use deodorants to mask the smell of a vape.” Quick to identify an opportunity to make a positive impact on adolescent health, MACSO developed a model that can differentiate between deodorant, smoke and vape, and has applied this to Piera’s aerosol sensors. Early trials have yielded excellent results.
Although Hwan claims to be “lackadaisical” about his career path, it’s clear that a passion for commercialisation – ignited early by that influential course at CIE – drives him. He emphatically rejects any assumption that the worlds of academia and entrepreneurship are mutually exclusive, citing an experience in Seattle that crystallised his thinking: “I asked some brilliant academics about the commercialisation aspect of the algorithms they were discussing, and they all looked at each other and laughed. That really bothered me – I thought, a lot of taxpayer money and resources are supporting you in making these algorithms, so why are you laughing about the idea of commercialisation? Commercialising your product is almost like a service to society.”
He’s equally passionate about the importance of keeping an open mind in terms of academic or career choices. “Don’t box yourself in,” he advises other researchers and doctoral students. “Don’t say ‘I’m a mathematician, so I do maths’, or ‘We’re academics, so we don’t think about commercialisation’. You’re all part of the same world, and to put up these false boundaries is just unproductive for everyone.”
For Hwan, one of the most rewarding aspects of his career journey has been realising the real-world relevance of his years of intense study. “I finished my PhD when I was 28,” he notes. “That’s a good chunk of your time learning. And it feels great to realise that that learning can go beyond academia, and watch it come into action. Algorithms I wrote, using knowledge I gained from my PhD, are now sitting on a farm in Spain. I just don’t know what can be more fulfilling than that.”
Applications now open for CIE’s Research to Innovation Hatchery – free for all researchers at the University of Auckland.
EMAIL
CIE@AUCKLAND.AC.NZ
POSTAL ADDRESS
THE UNIVERSITY OF AUCKLAND BUSINESS SCHOOL
PRIVATE BAG 92019, AUCKLAND
