A Data Scientist’s take on 2020.

Get in touch today

0141 225 0130

I, Joshua Smith (Client Acquisition Lead), recently interviewed Andrew Jones, Data Science professional with extensive experience within the Digital, Retail, Telecoms, Travel, and Gaming Industries.

Andrew is currently consulting on Machine Learning projects for PlayStation and is also the founder of analytics-link.com, a community hub for Data Science and Analytics specialists.

A man of multiple talents, he’s also recently written his first book “The Essential A.I. & Data Science Handbook for Recruitment“.

I asked Andrew what his thoughts were on Data science topics that he believes will gain in prominence during 2020 as well as his thoughts on the current state of the industry.

 JS: 5G is set to greatly increase how we interact with the world – an example of this would be further developments within augmented reality. Devices such as the Microsoft Hololens 2 would be a prime example of a piece of technology that could be implemented into the workspace in assisting users in a variety of daily tasks in multiple industries.

What practical elements do you think advancements in 5G could offer from a Data Science perspective?

AJ: 5G is going to have an impact eventually, but in 2020 I’d say that will be at the cutting edge of the field rather than for the majority. In saying that, the area I see it most likely to impact will be in small devices such as phones and “at the edge” computing devices. If 5G can transfer significantly more data, then those devices could, in theory, put more of their processing power into the cloud rather than housing it onboard. An example in practice would be something like a drone now having much more powerful and comprehensive object detection capabilities.

 JS: Deep Fakes are a great example of how Machine Learning can be seen in an easily understandable and accessible manner.

However, as novel as some of the examples have been, there is a serious risk of this technology being used to misinform and manipulate. What steps do you think the Data Science community will be taking to increase the detection of deep fakes?

AJ: We’re definitely seeing more and more Deep Fakes online, although so far this is mainly for entertainment (i.e. swapping actors faces). I’m yet to see them having a major negative impact as most of us worry they could – but only time will tell. In saying that, there is specific and dedicated work going on finding ways to detect Deep Fakes, including a Kaggle competition with a £1 million prize – so it is being taken seriously

 JS: What developments are you, personally, most excited about seeing come to fruition within the Data Science Industry? What does the future look like for you?

AJ: I guess seeing advancements in Object Detection, although we’re still so far away from human-level performance in this area. Language models took huge strides last year too, so it’ll be interesting to see how they continue to advance. They’re not quite human level, but 2020 could be the year they get there. The technology and hardware are always progressing too – seeing the bigger and more powerful GPUs come to market is always exciting to me. From a personal point of view, I’ve been working on a lot of smaller side projects (including an A.I. book for recruitment), as well as similar books for leaders and CEOs and for those looking to enter the Data Science field. I’d like to start building some online courses, as well as just build up my profile on LinkedIn.

 JS: What emerging Data Science techniques or tools are you keen to gain experience within 2020?

AJ: There are changes happening for sure – ones that we Data Scientists need to keep up with. The big players, namely AWS, Google, and Microsoft, are currently trying their best to lower the barriers of entry to A.I. buy automating everything. One aspect of this is AutoML which stands for “Automated Machine Learning”. This is the process of automating the application of Machine Learning to real-world problems, covering everything from selecting the best variables from the data, selecting the most appropriate algorithm for the task, and fine-tuning the parameters. It’s like A.I. is also coming for the A.I. jobs!

On a more positive note, I’m just going to keep developing my skills in Computer Vision, and Deep Learning in general. I’ve not really needed to get into NLP (Natural Language Processing) so if that opportunity comes up then I’d like to dig into it.

 JS: What advice would you give to aspiring Data Scientists for 2020?

AJ: Keep working on and playing around with Deep Learning, but don’t ever get too far away from your softer skills around commercial strategy and the customer. These are the core skills that A.I. won’t be able to automate quite as easily! Also, every so often takes time to go and revisit the fundamentals – things like probability and simpler models like Logistic Regression – I guarantee you’ll find that you have more to learn on these things that you think you’ve mastered!

JS: You have just released “The Essential A.I & Data Science Handbook for Recruitment” which is now available on Amazon. From your findings what immediate advice you could give to Talent Acquisition professionals who specialise in the Data Science field as we move into 2020?

AJ: There is no getting around the fact that Data Science and A.I. are fast-paced, rapidly growing, and highly technical fields. Keep learning and generally being curious about new concepts so you can stay close to the distinct needs of your clients and keep a rich understanding of the skills and experience of the candidates. Being able to connect those dots will continue to make you vital to both parties.

That’s what the book is all about really – I essentially lean away from mathematical notation and jargon, explaining and comparing all the key concepts, tools, and algorithms commonly listed on job descriptions and CVs in a lightweight, interesting and (hopefully!) fun manner.

I also spend time discussing the new ways that candidates are learning their skills and I’ve also provided some template questions for interviewing and screening – even for those who don’t see themselves as particularly technical. Hopefully, it’ll be another valuable tool in your toolbox!


Author: Joshua Smith