#136 Putting AI to Work: Delivering Actionable Insights with Machine Learning and Deep Learning with Dr. Alex Antic - Head of Data Science
Dr. Alex Antic is the Head of Data Science at the Software Innovation Institute. He has 17+ years of experience across different industries including Federal and State government, Insurance, Asset Management, Banking (Investment and Retail), Consulting and Academia. Alex has recently been recognised as one of the Top 10 Analytics Leaders in Australia by IAPA (Institute of Analytics Professionals of Australia).
He describes the early days of his career as a time for personal growth and honing in his technical skills. Now, his goal is to use data science and artificial intelligence for public good, and as a way to drive impact and change.
Alex is an enthusiast of the experimentation culture; he believes in the fail fast and fail cheap notion and how it is important for organizations to not be scared of failure, given that this can keep them from moving towards innovation.
As many other data scientists at certain points in their careers, Alex has struggled with a lack of support from management towards data-driven decisions. His tips for facing this are to avoid talking to them in technical language, instead approach and talk to them from the business perspective. This will help get your point across and show them the real value that data has to offer.
Stay tuned to know more about how Alex leverages data and Machine Learning/Deep Learning capabilities to deliver actionable insights.
Enjoy the show!
We speak about:
[0:40] How did you get to where you are today?
[6:00] What was the catalyst to move into Vol. 2 of your career?
[7:55] Tell us a little bit about the community-building side of Alex Antic.
[12:00] Tell us about the data science riddles you share on LinkedIn.
[16:27] Tell me about putting AI to work and your focus on delivering actionable insights using machine learning and deep learning.
[24:00] Do you have any examples that you can share with us about ways to get value from machine learning and AI.
[31:15] If it’s not being consistently applied or done this way in a data science team, where do you think pragmatism should come from?
[33:51] How can you influence positive change in your organization when management is not so supportive of data driven decisions.
[39:05] How do you prove to skeptical management that your work has value or has had an impact?
[42:15] How important is the business evangelist to the success of the data scientist?
[46:45] Do you have any thoughts on what is taught in universities in relation to the requirements in business for a data scientist?
[50:15] Do you see many opportunities for data science in rural areas or do you think this is a career that is only viable in cities?
[52:55] What has been your favorite project that you have worked on?
Resources:
Alex’s Blog: https://impartiallyderivative.com
Alex’s LinkedIn: https://www.linkedin.com/in/alexantic/
Alex’s Twitter: https://twitter.com/dralexantic
Software Innovation Institute on LinkedIn: https://www.linkedin.com/company/anu-software-innovation-institute/
Quotes:
"Can’t take out the human element when it comes to analytics. Analytics only gets you so far, you have to think about broader applications."
"People need to think about what problem they are solving and if it needs to be solved by a complex method."
"When you are trying to solve a problem, start with the simple solution first, and add complexity as you need to, as sometimes the non sexy elements will add value to the organisation, such as automating an excel file .. You don’t often or always need to go down the complex deep learning algorithm to extract value, as it will make it difficult to explain and difficult to validate, and simplistic is beautiful in many ways."
"There is always this trade off when building a highly accurate complex model with people less likely to adopt it or integrate in their workflow vs building a simplistic good model with decent accuracy that business can run with straight away and they find it easy to understand and work with."
"This notion of fail fast and fail cheap is imperative to data science culture, and organisations that lack that and are very risk averse, often struggle with doing really innovative work in data science. Having that culture of experimentation from top down is imperative, and organisations who have that have excelled in innovation whereas organisations that are more fearful and scared of failures, have been really slow movers towards innovation."
"Universities should be able to integrate their transitional research, the work they are doing with that of business by understanding business needs and how business wants results often quickly in a way so they can develop an IP on the things they are doing and scale it. And businesses while working with universities should understand that despite universities are doing things for them, they also want to be able to use it for other purposes such as publishing or for their own research purposes."
We are now on YouTube! Watch the episode here: https://youtu.be/EiI6dU8PSqs
And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. It really helps new data scientists find us.
Thank you so much, and enjoy the show!