Michael Berns is Director for AI & FinTech at PwC. Michael has a broad background across blue chip names such as Morgan Stanley & Moody’s as well as a range of smaller innovative AI Firms. Michael has been a Mentor and Judge for organizations like Startup Bootcamp, Virgin Money Startup, Cocoon Network, Level 39, MIT Inclusive Innovation Competition and the United Nations World Food Program. Michael is als a guest lecturer at London Business School and Mannheim Business School. Michael holds an Executive MBA from London Business School. s an

In this episode we revisit AI in Finance and Banking and we are busting some myths about common misconceptions people have about it’s adoption. We also get into the difference in AI adoption across several regions across the world. We talk about regulation, use cases, and much more. Michael also gives some advice on getting your career path right and resources on how to start with AI.

Links

https://www.linkedin.com/in/aithoughtleader/

https://twitter.com/AIThoughtLeader?s=20

https://www.forbes.com/sites/antoinegara/2017/05/23/wall-streets-robocop-how-digital-reasoning-cracked-the-code-on-trader-talk/?sh=31318b7a3520

https://www.pwc.de/de/finanzdienstleistungen/kuenstliche-intelligenz-im-finanzsektor.html

https://www.pwc.de/de/future-of-finance/how-mature-is-ai-adoption-in-financial-services.pdf

Contact

If you want to get in touch: contact@thewallstreetlab.com

We look forward to your mail and will do our best to reply.

If you want to reach out to us personally, here are our LinkedIn profiles, please mention the podcast.

https://www.linkedin.com/in/andreasvonhirschhausen/

As always, please do not forget to take 17 seconds to leave us a 5-star review on Apple Podcasts or wherever you get your podcasts from.

Be well

Andy

EPISODE 79

[INTRODUCTION]

[00:00:03] ANNOUNCER: Welcome to The Wall Street Lab podcast, where we interview top financial professionals and deconstruct their practices to give you an insider look into the world of finance.

[00:00:23] AVH: Hello, and welcome to another episode of The Wall Street Lab podcast. With me today is Michael Berns. Michael is director for AI & FinTech at PWC Germany. Michael has a broad background across blue chip names such as Morgan Stanley, Moody’s, but also of a smaller and more innovative AI Firms. Michael has been a mentor and judge for organizations like Startup Bootcamp, Virgin Money Startup, Cocoon Network, Level 39 MIT Inclusive Innovation Competition and the United Nations World Food Program. Michael is also a guest lecturer at London Business School and Mannheim Business School. As a educational background, Michael holds an executive MBA from London Business School. Michael, pleasure to have you here.

[00:01:13] MB: Andy, hi. Pleasure to be here.

[00:01:16] AVH: It’s a pleasure. Let’s start with a report that you offered in 2020. In the introduction, you quoted Google CEO, Sundar Pichai, to be saying, AI is probably the most important thing humanity has ever worked on. Then you quoted Stephen Hawking to announce that the development of a full AI could spell the end of the human race. Let’s start there. Are you a techno optimist or techno pessimist?

[00:01:51] MB: I guess, techno optimist. We did have a PWC test on the website where you could select, and I think it was done globally. I was with the Chinese view of AI and some elevated from. It’s for me, the reason is why I am an optimist is that they’ve implemented AI solutions in a number of clients or last eight years. I could really see the beginning of an actual use case to do all the way using it across the board, from tracking down the bad guys in banks to building use cases for preventing human trafficking. So using AI for good, and I think there’s so much potential for it to be used for the right things. But there’s also regulation and frameworks that we need to do put in place to make sure that it’s used responsibly.

[00:02:45] AVH: Do you also very, a big advocate of AI for good finance, for good, but let’s start maybe with the base. We recently, we talked in a couple of times in the podcast about AI, but let’s maybe define some terms, because there’s a lot of confusion between things, just use case like neural networks, what is data science? What neural networks was actual AI? What differentiates from machine learning? Then for example NLP. Can you just give listeners a really broad overview of what we going to talk about today, and maybe also where we stand in terms of development standpoint.

[00:03:29] MB: So even in between the different experts, the terms are somewhat used into change of place, so you will always find people who will put rule based in mentor, artificial intelligence, then machine learning, of course, deep learning and so on. So when the border is really is still very fluffy. I can define it for me. I can rightly say that some things that we have delivered, a human being even more for particular intelligence and very switched on would not be able to deliver. So we are using the machines ability to see endless number of details in parallel, something that human being cannot do. So mostly, a lot of the cases that I’ve developed are around understanding language at a very nuanced level. 

So if you give the human a PDF to read with 100 pages or longer email trail with different actors in a bank. What you find is that once you’re reached item 20 or 30, at some point, human brains which is off the computer and the solution using machine learning, maybe natural language programming, other technologies is able to pick out the irrelevant facts and present them back. So when I’m building a solution, it’s usually to empower the human to do the financial text but I’m doing a lot of the texts for relevance, the loves of prioritization on before, and that is helping the employee or the caseworker or the executive to make decisions that otherwise he wouldn’t have enough information on or where he would be overwhelmed with information. 

So in short, if you go back to 1950s, you could say that a system that can think for itself or make decisions can be classified as artificial intelligence that were the beginning of the definition, but of course from the 50s onwards, a lot of the mechanisms and a lot of the technology has been built on. Now it’s up to ourselves to say where it draws the line. What you can say is, and this is because he quoted, Hawking earlier to say that once we get to the artificial journal intelligence, which is curiously always some 20 years out, yeah, so might never happen. Then you will have a system which could replace the human for all kinds of questions, all kinds of situations. 

We’re not there yet, but we have different fields like financial services, compliance area or healthcare, detection of illness or finding alerts and in a wide number of cases, where the machine can just be far superior of seeing patterns, putting those together in a readable form, transforming it and making it possible. But as with in medicine, the best solutions are the ones where you combine the human insight and experience, but make it so you take out their routine for them, make them to focus on the other cases. The combined solutions usually the one that has the highest accuracy and precision. I don’t see that to change or there’s no need to automate everything out there.

[00:06:57] AVH: The more specific the use case and the more routine and the more repetitive something is, the better is probably AI for it versus a general AI that it’s very so for me, it’s a very fluffy term. So maybe you could help me and the listeners to understand, what would be a general AI for example be?

[00:07:20] MB: We can say curing, defining with the Turing test, so a system that could give human-like responses, and you wouldn’t be able to distinguish anymore, whether it’s a person or a bot. I think, we’re nearly there, when you think about some of the big language models being published over the last year or two, and constantly being updated. You have solutions where then have all the knowledge in the world, based on Wikipedia, and so on. They can play it back to you in a scenario where you asked a question and a solution response.

Then, it’s also very contextual, depending on what settings, then write a poem and give some answer back, and so on. This is technically possible. What I’m saying is that so far, we have this narrow intelligence where we build solutions that are tailored to one use case. So find front running in a bank or find a potential cancer and X-ray and doctor’s notes and so on, and it’s only in to this. We haven’t built a system that has the human ability to just learn without much context in all situations. Maybe for some games that’s true and for some other things, but it’s not quite childhood way, where you can just say, I throw it in a different situation and it will intuitively do the right thing. That’s not where we are yet. I doubt we will be there that soon. But just to generate the answer or I have a very narrow use case. Sure the machine can be the human in this narrow perspective.

[00:08:55] AVH: But then where is this fear of AI coming from? Is the fear coming from one day, maybe not in 20 years, maybe 30 or 1,500 years, there will be a general artificial intelligence that can adopt to everything we humans do, just better and then we won’t be needed anymore. Everything that you mentioned so far, doesn’t sound very scary, right? It’s just sounds very, very normal. But where is this say the negativism around this coming from?

[00:09:26] MB: It’s a lack of understanding. So let’s say when you have been in this field for eight years, you exchange a lot of global experts in this field. We’ve been on all kinds of panels, you’ve discussed some use cases across the world, including China, with social scoring, you’ve built some solutions in Asia and US. You get a sense that it’s like the arrival of the Internet and that’s what I usually compared with initially in the 90s. People were also maybe outraged about some of the information that is out there. 

People find Instructions on how to build a gun or can form organizations or some do all kinds of things and when we realized, well, without the internet, we would not be as knowledgeable or have information at our fingertips today, that would be a good chunk missing in our ability to assemble information whether it’s from YouTube or whatever source you have it. Everybody today except that this is a given this is there to stay. But it back in the 90s. I remember they were not sure if it’s a force for good. They were not sure if this chaos and this allowing people to put out anything is a good thing. 

There was a lot of debate around this. In my earlier years talking about the subject on many panels, I used to compare artificial intelligence with atomic power. Of course, I realized that this has got more of a negative connotation. What still rings true and that goes for the internet as well that all these arrivals of seriously disruptive technology, it’s still in terms of who is using it, for what cause. I feel that it’s a task for the government and for companies to take the employees with them on a journey for the government to not create unnecessarily fear of the subject, but more say that this is something that is happening around the world anyways. It’s our choice now whether we want to play in this or withdraw, but it’s already done. 

When you speak to a lot of my contacts in China that I’ve built over the years doing business development in Asia, they don’t necessarily see it as a negative thing to have social scoring. They see it as a equation where they have some benefits out of it. They share more of themselves and they have different products recommended or offered to them. They have faster shopping. They feel that they are safer in the neighborhood, if they have social scoring. We can look from our German and western perspective and say, “Oh, but there’s no personal freedom and this is not democracy and this and this.” But it’s unless you really talk to a lot of people on the ground. I feel that there’s always a cultural gap that you can’t overcome in terms of understanding both sides of that story. 

I’ve seen also in China, how artificial intelligence can overcome supply shortages and other things where I think the scaling and the infrastructure, many things wouldn’t cope without some of these mechanisms being automated. Yeah, so when you talk to different market participants, including China, you find out that for some of them, it’s also necessary for scaling and as well as the functioning of the economy to automate a lot of their supply chain, their resources and on using AI. So it’s used across a wide number of use cases. There is no reason to be afraid. But  any technology, it depends on what to do with it.

[00:13:17] AVH: I asked about the bad things, about the evil general AI, but let’s talk about the actual, the good thing, the positive things that can come out of it, maybe also in terms of the financial industry. As you know, our audience are most mostly financial professionals. What use cases do you see for AI machine learning to really have a positive impact on the financial industry?

[00:13:44] MB: I’ve been part of financial services pretty much since beginning of my career, somewhere around 2003. First working for the regulator and more years and later Morgan Stanley going through the financial crisis. So with the arrival of the financial crisis, you saw a complete change in perspective, in terms of what business the banks are interested in, what the regulator wants them to do, and what the customers expects their banks to do. That means that there was a significant shift, of course, towards the FinTechs or to challenger banks on, along with a loss of trust. 

Then, of course, these from my UK, American and global perspective, came with the arrival of some massive fines. So when I say massive, I talk from 200 million per bank upwards to several billion and this kind of effect the – the after effects of the financial crisis where some of the things that we probably have seen in the market before, like the manipulation of the labor or ethics, manipulation or front running has been going on for a substantial amount of time, but it’s only now that the new technology allows to go deeper and catch some of the perpetrators. 

So what it does, from my perspective, having built some of these early use cases in these areas, it gives the financial services institution a another tool to be able to say, “Okay, I can control better what’s going on in house, I have a better view of what else might be hitting me.” Yes, it’s going towards some surveillance, but if the choices are, do I have some more employees, some traders violating some rules, manipulating global benchmarks, where products worth several trillions are tied to or I rather stay away from surveillance and I go as usual. Then the choice has been made very clear for the UK and American banks and this is more also acceptable in Germany now. 

By analyzing language at scale, so ingesting the entire communication within some of those banks, so emails, voice, the messengers and on the different channels, and then trying to find the needle in the haystack where some people collude, where there are certain patterns of giving each other credits after something has been perpetrated or some indications of using code language, all of these things are not something that the AI automatically can find. But you can train a model to be robust enough to show you when it’s certain with a certain percentage. 

To come back earlier to your question, where you asked, okay, where’s the cutoff or so I mean, at the base of it, it’s all advanced statistics, now we’re talking about the chance that this chain of chance would contain something suspicious that should be investigated by second line of defense, and then by legal. So you’re working with training the model working with some percentages here, trying to cut down the noise. So rather than having 300 alerts for the team to investigate you, you use the solution to make it 30 make those investigations count. The other side of the coin is if you can work with a lot of communication within the firm and a lot of the things ingested from outside the news or social media, you can also drive product development and research. 

So some of the use cases I’ve built over the years, were more in the research arena or market insights in terms of what is important to our clients. But you can also give clients another way to interact with you be it as a chatbot, voice solution or digital channel where you might use the system to propose some investment options or relevant products to them. Maybe the acceptance of those suggestion is not quite as high as yet, as in Asia, but for a lot of China it’s if the customer feels there’s something relevant being put in front of them for the right price, they see it as time benefit in terms of having information available as their fingertips and making the right choices. For them some of the applications and platforms that are out there are more an extension of the search functionality in the web is just more proactive approach. 

For me, in short the ability to use artificial intelligence in financial services gives scale, gives opportunities for new companies, new solutions, digital banking, mobile banking, some viral campaigns, some of the challenges. It also gives the existing banks an opportunity to tie down their compliance systems and fraud detection, anti-money laundering and those areas. If you use it in the right way, it can give a lot more insights than before and drive a lot more client decisions and discussions as well.

[00:19:15] AVH: You mentioned two very different use cases that we discussed on the podcast fairly recently with Daniel Faggella, where he basically says a lot of the banks they tout in social media, their chatbot use cases, just because it sounds cool, it seems fun, but actually 90% are now making this number up, but a very high percentage of funding actually goes to what the later use cases you described. Anti-money laundering, fraud detection, is this something you see us well, would you second that or do you have another opinion on it?

[00:19:54] MB: This chatbot topic has been around for a long time in the 80s you had this Eliza thing which was pretending to be human and displaying back some empathic responses. But even for me, even in my time in artificial intelligence that has been with me throughout the whole period. Now, when you look at some of the virtual assistants available in the Middle East, for some of those banks or in Singapore and in some other jurisdictions, you’re blown away how much tailored the experience is for the clients, how much the appearance of the whole solution is, how advanced and this is. I haven’t seen anything this in Germany. 

Yet again, one of the things I’m also doing is developing a chatbot and why because in some ways, this is a very, let’s say, easy building block for people to understand that there is an intelligence providing insights in terms of hopefully not just a question answer system, but I can open account, I can look at my balance, I can order products, I can do a lot more than just firing questions back and forward, I can have a hybrid handover and so on. So for me, I haven’t seen so much advanced chatbot development by German banks yet, yeah. But of course, they’re all looking into this and have been for some time. 

I think the money is spent and that’s a result of the financial crisis and the regulatory crackdown. Afterwards, the money is spent protecting the banks building compliance use cases due to the budget available in those areas. So it’s one thing that you have to do for the regulator and for you to keep running. The other thing is, I wish this is something we also put in the study that you mentioned earlier. I wish there was more development it is in US, in terms of how do we build a new business model for a separate entity? Or how do we come up with an entirely new way of engaging our clients, but most of the development here seems to be more defensive, driven by the burning platform on the requirements around this. So there’s a bit of a different maturity and stage depending on what type of products.

[00:22:14] AVH: Where on the map, would you say AI adoption is actually implemented? Or where’s it actually, where the banks walk to talk and where it’s rather showcasing here or there. But it’s not really something done. It’s a small pilot project. I don’t know, just saying Germany, it did just mostly do small pilot project, but in Singapore, in Asia, in the Middle East, they really do get out the big guns and really implement AI use cases. Do you see a difference in where do you think each bigger, larger lead legislation is in India adoption of AI?

[00:22:58] MB: The main driver there is the regulatory angle. Till maybe half a year ago, I would have said and you know and financial, and WeChat, and all the things around it, to build their own ecosystem in terms of payments, getting the payment data, creating profiles and so on. The reaction by Chinese government and the regulators there has shown that there are also limits to where they can go. At the same time, then you have the US where, also till recently, it was you had to get approval for your FinTech solution or your challenger banks, in some cases for each individual state, so it was very difficult to scale. This has now been made easier under new regulations to address this. 

In the US, I didn’t see it so much in the pure play banks. I mean, maybe were some challenger banks that were founded out of pure play banks, such from the large investment banks for example. But also I see it more in wealth management, asset management, robot advisors, those are the unicorns ones that have gone big. Then you look at Europe, certainly, also some of the challenger banks use it. I don’t want to name individuals here, but I can just tell you in level 39, that you mentioned earlier, we had some of this largest European unicorns on the same floor in terms of payments and challenger banking providers. 

They’re used it more for building a different bar engagement model with new clients, building chatbots for onboarding for helping them, but also the integrations. So when you look at a lot of those new companies being built, they’re not profitable with the standard bank account. They’re profitable, because they find an easy way and automatic way to integrate with a wider ecosystem of products that they can offer. This for them is then the game, so very low margins. But if you find a way to scale with AI to grow fast enough to use it to reach a wide number of people to put the right products in front of them, even if they are not yours, but they’re from another provider, you can still be reaching profitability. 

I think this is where we will see maybe within a year past COVID if their economy is turning back to normal, and the funding at some point is not as freely available, which ones are these survivors. When you look back at the study published last year, some of the statements made by CDOs and existing large banks in Germany and also that they cannot exist in the same form if they’re not fully adopting AI now. So there has been a lot more investment over the last two years. Also from Germany, I would say, I’ve seen some of those teams go up from like 70 people to 300, 400 people. So if they do take the subject, very serious, but they are still, let’s say they haven’t seen the same profitability like some of the UK or American banks have seen come back after the financial crisis. So it’s not been a investment case for them. It’s been more carefully choosing where to use it on. 

If you look at the annual spend of some of the largest American investment banks, it’s several billions spend in big data architecture and then building up teams around them getting the best people from Imperial College, Oxbridge,  and so on in the UK, but also then Stanford and elsewhere, right. Billions are spent over there, but after a lot of infrastructure investment, and a lot of data management investment before, I think we are more now at the infrastructure investments for some of the smaller houses in Germany and Switzerland and Austria. 

[00:26:59] AVH: Do you think that’s a factor in the outperformance or, for example, US banks versus European banks, that they have built better earlier use cases in FinTech Financial Technology, AI, and now that they’re reaping the rewards, or is it something unrelated, because they have been investing much more into technology?

[00:27:29] MB: So it’s a tricky question, I think I mean, the regulator in terms of pre financial prices will start in the long legs in the US, so they didn’t have the same rules. But being inside Morgan Stanley during financial crisis, where you’re CEO has to go on a weekly town hall to tell you, these rumors are not true. No, we haven’t approached Citibank, because we can’t go without them. Then one year later, you find out the bank took 70 billion behind the scenes to survive. They have been propped up, but they have with their upside investment approach, we need to go into certain areas. They’re found new areas of profitability. Then I look at the investment banks, some of them have turned themselves into bank holding companies. Some of them have invested heavily and wealth management, solidified their client base, others have built challenger banks to work internationally. 

So there have been a lot more active. I feel the AI is only part of this. I think it did feel even the financial crisis that Morgan Stanley where I was, it did feel at times, like half of the bank was the development shop. Yeah, we were building products and solutions for our clients, which were their hedge fund at the time, so for prime brokerage. We were building solutions for other market participants to get a bigger market share. So it was quite digital, and it was quite ahead of the curve. I think after the financial crisis, they realized that they have to really go back on those big data infrastructure investments and big data. 

They automate a lot of the product literature, the funding flows, the terms and conditions, the collateral. All of these are now automated to a degree where they see trading opportunities that nobody else can see, as quickly. So brings us to other topics like low latency trading on. But it also brings us to areas where they decided to broaden out. I feel the crisis in Europe was a lot of the participants including Germany more, “Oh, we have over indulge in these things, so let’s reduce down here and oh, we need to cut off this.” 

There was a lot more Chronic questioning whether they need certain units and no one was more questioning of after buzzer free and then ongoing from there, whether it makes sense from a risk perspective to keep FX trading and so on. So there has been a lot more shifting in terms of their focus in terms of prop trading, equity trading and so on.

[00:30:25] AVH: You mentioned earlier or we come back to this topic over and over again, to have regulation. What would you wish, from a regulator to set up a good environment, so AI can be a force for good? 

[00:30:40] MB: When I arrived back in Germany after a long, long time away, four years ago. Then realizing that GDPR, and then the German privacy laws, I feel they were made at the time when there was no understanding of the importance of artificial intelligence. Now, Europe has gone ahead and drafted some other regulation, some rules in terms of these responsible AI use cases and these are not in need to look out for biases in the data and so on. These our tenants are going in the right direction and they show more understanding than before. I think what’s still missing is alignment of GDPR and awareness of artificial intelligence. 

This is gone to where it’s not practical and when I think about some of the surveillance and conduct use cases that are built in banks, it shouldn’t be a case where for the bank to decide, “Okay, do we want to survive? Do we can we have another $2 billion fine? Or can we convince our workers council on to get one of those systems in so we are more in control?” So there needs to be, and that has shifted in US long time ago, and awareness that as you join a financial services institution, as a trader, as a private banker, whatever it is, your communication and your responsibilities to the client, you have full visibility and responsibility to that. 

The bank in order to protect itself and continues it’s alive, needs to be able to access that communication and be able to act within law to prevent manipulation, prevent fraud cases and so on. There shouldn’t be debate whether communication is then private and should be protected. It cannot be private. The higher purpose is that things stay within the law and there’s no fraud case of billions of dollars and a lot of the investors, the bank, and others use their money, right. There needs to be a discussion around the value of data and transparency of what data we share. Equally, there needs to be one around responsibility and being able as financial services institution to control better what’s going on. 

You can realize it’s a bit of a philosophical thing, because I feel, we are on the right path. There’s still a lack of understanding to fully implement this. I think we have enough of a niche in Europe to find something between the American and the Chinese approach. But we do need to speed up our efforts around guy eggs and building a federated infrastructure and so on. There is a niche, but it’s getting smaller by the month by the year on.

[00:33:48] AVH: Now, you mentioned earlier that US banks especially invest a lot of money into hiring the best people, hiring a from Imperial College London, Stanford. You were to start again, in the financial industry, maybe in general financial industry, or we’re interested in artificial intelligence FinTech for a banking use case. What kind of education would you get? What practice would you try to get to get a good and solid background to then be a one of those top candidates that is definitely gets a seat on Goldman Sachs’s quantum or AI Data Scientists Team?

[00:34:42] MB: I worked with some of those individuals. Back in London, in the early years, it was still difficult to find people have a lot of experience in machine learning and so on. But even in Germany now. There’s more than 200 programs at universities where you can study this. I feel that this new way of working, it doesn’t only require machine learning and data scientists. It acquires in more agile mindset more of a solution building in a dynamic setup way. There’s room for product owners that there’s room for Scrum Masters in terms of building this team out, but also from a development side. 

If you want to make it to some of the largest, both bulge bracket banks, of course, then if you have an early interest in coding, and I’ve seen some of those candidates go through successfully, some of them started coding with 10 years, 13 years, I always felt they had to take part in some of these public competitions be in cargo or other means and online and in different data science championships, tests how far they’ve come along. Then some of them join those teams. They realize the skills that I had that were really good for coding on myself and quickly building a prototype, but it’s maybe not quite enough yet to build a product professionally from enterprise readiness in terms of for clients. 

So they still very steep development path for them bringing even the best coding skills that you can learn at one of the leading schools to cooperate in working on something with eight, nine elopers additionally, in the team in a Scrum or in a Nexus or with a lot of different Scrum teams to contribute to something that is growing dynamically out of the effort of several teams. So then comes more coordination program. So it is pretty much as in the past, tried to go to one of the top five schools. Try to broaden your skill set from not just business or not just coding but have a sense for the markets early on. 

Personally, I think it helped me a lot in my long standing interviews back then, that I had started investing in the stock market, from age 18, that I couldn’t really tell what was going on with Dow Jones and all the different indices that I had an idea about financial products. A natural interest, really, that these were all things that I would spend my free time in learning. Then I did some back then VBA coding on the side me some rudimentary skills. But it was more my interest in this combined with solid education combined with maybe some of the charities I support it in free time mentoring startups and so on. 

It’s a combination of factors. I feel you should not bet it on one card just trying to get into one of them. I think it’s more important that you work on financial products that you challenge yourself, that you have a go to attitude and a mindset to be successful, because what I also realized in the financial crisis, that all these places are great to get into and you will feel you have accomplished something, but it’s the what happens after, what does make you successful, staying in there, no matter what happens with the whole market and the surrounding. That’s the the bigger challenge. 

They’re getting in is one, but remaining relevant to your clients and to your marketplace. So it’s constantly learning. It’s a very dynamic area when we talk about artificial intelligence. There’s another big language model down the line. There’s another tool that Google or MSN publishes. There’s another way to augment or to label later. Something that makes you just a little bit faster. You’re meant to stay on top of this, in addition to your learning and your programming. It’s not another easy task. The good people are really in very high demand globally. They can pretty much command what they like in terms of job titles, remote or salaries in some cases.

[00:39:15] AVH: Perfect Michael, it’s been really insightful. It’s always hard to keep to stay relevant in the in the space, but I think with general data science programming, especially with a specialization on AI. If you love what you do, you can do stay relevant. Now before we wrap up. Do you have any thing that you want to address our audience with? Any tips, any comments, any way they anything that you want them to think about?

[00:39:47] MB: For anyone who’s really interested in artificial intelligence of course, there are many books out. I’ve also contributed to one, but you can, of course that also with podcast, this one and in some very specific ones, but it’s like you will understand the business sense at some point in terms of what are the probably the likely use cases, where can you create value. That’s great. If you want to go deeper, then I think at some point, you have to really decide whether you do some Python tutorial or we take some of the Stanford classes on AI, where it goes into the mathematics and the things behind it. 

I think, what I can see and people applying to us, there’s a big difference between candidates who spend some of their free time coding and really naturally try to get into the architecture of different solutions or people who’ve just put Python on the CV. Good at finding things on Google or copying it from GitLab and so on, and then just plugging the code in. There’s quite a difference. I would rather take the candidate that has a natural interest and doesn’t follow the hype, and this, maybe more of a good coder who’s also capable to ride in Pi from rather than the other way around. 

This is just some career advice. But I think there’s a lot of sources out there. I think you just need to find your area of specialty or your interest to get the motivation to build something and the way that companies are hiring and that field is also changing. So you see a lot more of these hackathons or the individual introductions. You see a lot more CVs with code repositories on them, where candidates share what they’ve built, and colleagues actually looking into this. So there is a change coming in terms of what is required upfront to get into an interview. Of course, there’s still a lack of skills, no doubt, but there’s a different transparency in terms of what somebody has done or can actually deliver before you even talk to them, which wasn’t there before. 

[00:42:07] AVH: I think it’s probably good to get into space to actually show you work to be active and not to have to the theoretical credentials there, but have the practical credentials and actually show that you’re capable of doing, the use case you don’t more than the theory. 

[00:42:24] MB: Yeah, I think when you show that you’ve got a long term interest in this, that you have developed of being part of a number of projects or solutions in that year, that you’ve come to contributed maybe to some open source things. All these shows interest, in it’s in the end it’s your ability to learn, because that area is so dynamic, hardly anything else. So it’s not what it’s how fast you can assimilate and learn. Then it’s this times, times your motivation, how much you want to deliver and that is ultimately the format for success. 

Of course, we all would love to have and hire the social data scientist or have somebody with empathy and show skills as well as the coding skills, but maybe it’s a cliché, but it’s sometimes to find the right balance there is can still be hard to go the business side and who wants to deal with client rather than data problems. This is the next step, just something to keep in mind as you build your profile and your abilities in this field. See how much you like dealing with clients and direct exchange as well as you like building solutions and thinking about big problems there.

[00:43:49] AVH: Awesome, Michael. Thanks so much for coming on the show and sharing your insights with us. I wish you a great weekend.

[00:43:56] MB: Thank you so much Andy. Great weekend to you too. It’s been a pleasure being on here. Looking forward to the recording.

[OUTRO]

[00:44:04] ANNOUNCER: Thank you for listening to The Wall Street Lab podcast, for the show notes and much more. Visit us at www. thewallstreetlab.com. To see what we’re up to before anyone else, subscribe to our newsletter on our website and follow us on Facebook and Twitter.

[DISCLAIMER]

Information contained in this podcast constitutes the opinions of individuals and should not be treated as investment, tax, financial or legal advice. We take no responsibility for the accuracy of any statements made in this podcast. This podcast is for informational and educational purposes only. It does not contain an offer to sell or buy any financial products and should not be treated as advertisement for such. Any copying, distribution or production of this podcast without the prior permission of the creators of the podcast is strictly prohibited.

[END]

Published On: December 15th, 2021 /