We caught up with Josh Greenslade, Machine Learning Engineer, to find out about what the life of a machine learning expert at ONYX InSight is like, and how he made the transition from astrophysics to renewable energy.
Hi Josh, could you start by telling us a bit about yourself and your career so far?
After a master’s degree in Astrophysics at the University of Nottingham, I did a PhD in Astrophysics at Imperial. I studied the formation of dusty starburst galaxies, then did a 6-month post doc there, working on, amongst other things, 3D printed planets and educational outreach talks. Then, I moved to machine learning at ONYX InSight.
When did your interest in engineering begin?
It was really when I moved to ONYX InSight, actually. One thing I found was that many of the skills and methods were very transferable. Detecting faults and anomalous data, statistical distribution and mathematic models – it’s working with numbers. If a main bearing fails in a wind turbine, the diagnostic process is broadly similar to sorting data in astrophysics.
What sparked the move from stars to wind energy?
I’ve always been interested in the energy sector. The industry will be very important over the next 10-20 years, and we are likely to see a dramatic shift in how it is run. There hasn’t been change so rapid in energy since the move to coal. Previously, there has been industrial inertia and a reluctance to change – but now, people are open to adopting new and exciting technologies and ways of working.
Astronomy was interesting but often too abstract. As a machine learning engineer with ONYX InSight, I can tackle practical problems and create benefits for everyday life. The move was from a more academic setting to one where I feel I have more utility.
Can you give us an idea of what might be included in your day to day responsibilities?
A typical day involves collecting and cleaning data, standardising it and removing anomalies. Then I can start to get a feel for what the data can tell us. Typically, through a process of data questioning, involving mathematical queries and input from engineering experts to begin problem solving.
From here I can bring in and tweak machine learning models, testing and training new models on historical data. Then, once I have got a solution, I will write the finding up in an easy to understand report. There’s often a fair amount of presentations and talks on my field which I give as well, on top of the data analysis.
What would you consider to be the biggest challenges in your job?
Data standardisation can be a challenge. With multiple turbines and wind farms with varying levels of digitisation, and sensors, making the call on what data to use is down to experience. The data must fit well in the real world.
Typical data problems include a lack of historical data or relevant information, or limits on what the data covers. Equally, it might not be stored well, so we might need to go back to the customer to ask for more information. In some instances it can be difficult to know if a problems is solvable using the data as it is handed over.
What motivates you on a daily basis?
I really enjoy trying new techniques and finding a novel solution to a problem. After you invest 20 hours in a problem, the breakthrough moment is like striking gold. It’s hard work but very fulfilling.
For you, what are the most exciting developments in renewables and in your field?
The trend towards bigger renewable energy companies and bigger wind farms is something I expect to see more of. That said, on wind farms we are seeing less individual turbines – but bigger, more powerful turbines. With artificial intelligence, I think the industry is getting to the point where it understands where the technology works best and expectations are realistic. In five years, the conversation will be very different. As the technology becomes fully mature and standardised, machine learning will become more seamlessly ingrained into the industry.
What would you say to young people looking to get into renewable engineering?
Read. My colleagues and I all come from very different backgrounds and disciplines – engineering is diverse, and a company like ONYX InSight has many roles, from sales and marketing, to software, consultancy, and heavy engineering. One thing we all agreed on, is that reading is important. Be interested and curious about the world.
Finally, if you had to pick one, who is your engineering hero?
I wouldn’t want to single out one person. I have enormous respect for the many engineers working tirelessly, anonymously, and often without credit, to ensure that basic things work. We often only realise their importance when things go wrong. These are the people who ensure your tap supplies water in the morning, or your internet ‘just works’, and there’s power to your house. These are the people who create and maintain our medical, food and transport infrastructure. So, thanks to all the unnoticed in that regard.