In this podcast episode, Dr. Beverly Wright, Vice President of Data Science & AI at Wavicle Data Solutions, and Akhil Mahajan, Director of Data Science at Procter & Gamble, explore the transformative effects of AI on data teams, discussing crucial skills data teams require, the challenges of inherent biases, and the need for strong core competencies to create high-performing teams. Whether you’re a data professional or an AI enthusiast, this conversation provides key insights into how AI can be leveraged to develop and improve your data team.
Speaker details:
Watch the full podcast here or keep scrolling to read a transcript of the discussion between Beverly and Akhil:
Beverly: Hello, I’m Dr. Beverly Wright, and welcome to TAG Data Talk. With us today, we have Akhil Mahajan, Director of Data Science at Procter & Gamble, and we’re talking about how AI impacts the way we develop and grow data teams. Welcome to the show, Akhil.
Akhil: Hi Beverly, thank you for having me.
Beverly: Yeah, my pleasure. So good to see you here in Cincinnati. Tell us a little bit about yourself.
Akhil: I’ve been fortunate enough to work with data and AI for the last 20 plus years. I have worked in a lot of different domains in automotive, manufacturing, healthcare, sports, and now at CPG. I’ve had the chance to work with a lot of data, collaborate with great people, and gain insights into various industries. In my current role, I lead data analytics for purchasing and finance at P&G supporting our users with their needs from a data analytics perspective. That’s not the only practice, we have a lot more data analytics teams.
Beverly: So you have internal clients?
Akhil: All of them are internal clients.
Beverly: Would you consider yourself vertical agnostic or more CPG specific?
Akhil: I’m vertical agnostic.
Beverly: Seems like you have a very diverse background. And for your education, were you formally an engineer?
Akhil: My bachelor’s was in mechanical engineering. But I always loved computers, so when I came to the US, I switched my master’s to distributed databases.
Beverly: Well, good. I’m glad to have you. We’re talking about teams today. Can you tell us a little bit about your tenure? When you talk about teams, not just the data and analytics or data science or AI team, but just in general, can you describe a little bit about that tenure?
Akhil: P&G is a great company. I’ve been with P&G for about four years, but I know people who have been here for 25 or 30 years. In my 20 plus year journey, I can’t even imagine working in one place for 20 plus years. So, it’s a testament to P&G on how well they take care of their people.
Beverly: Can you imagine the changes they’ve seen at one company in 25 years? It’s hard to imagine. I know we were talking before about how they don’t want you to become stale in one spot. So, talk us through what that looks like. Do they encourage you to move around or what?
Akhil: One of the best things about P&G is they allow you to move every two to four years, depending upon your role and seniority. So that way you get to work with almost every part of the company and you get to learn more about the company. In my current role, I’m finance and purchasing. In my next role, I might be looking to do something else like supply chain or manufacturing. That’s one great thing where I see why people spend 30 years of their career here because you never get bored because every three years it is a different challenge, which is amazing.
Beverly: Right now, you’re the Director of Data Science for finance and purchasing. But tomorrow, it could be people analytics or HR or something. Still maybe Director of Data Science, but in a different business unit. Do you think that’s the secret sauce?
Akhil: Absolutely because having a 25-to-30-year long career means you are being challenged again and again. I’ve been here for four years now, and I’m excited about the next challenge because I know I’ll be doing something different. That excites most people; it excites me for sure.
Beverly: Even though you’re having to relearn and get to know new teams, it sounds like it’s still worth it because it’s really enforcing people to stay.
Akhil: Absolutely. During meetings, I often encounter individuals who have worked together for nearly two decades. That’s amazing to hear, right? It’s good because there’s camaraderie among people; they know each other, and it becomes easier to work, communicate, and coordinate.
Beverly: We’re talking about how AI is going to modify the way we develop and grow our teams. What are the traditional skills required for data, analytics, and AI, and how is AI modifying these skills?
Akhil: If you look at a data team, they are made up of different roles, people, and personas. So, we all come together to deliver something for the consumer. Most of the traditional roles like data engineer, data analyst, BI developer know about SQL. Now with the latest data engineering trends, you need a lot of Python and PySpark. From a data analyst and BI developer side, you need to know how to do data visualization and tell a story. These are technical roles, but on the other hand, you also need to know how to present and convince people. This is something you need to use from a data perspective. With the help of AI, one of the biggest things I’m seeing is that now we can handle a lot of unstructured data.
Since I’m working in a purchasing domain, I deal with a lot of contracts. P&G is a huge company where money is spent to sign a lot of contracts. Now using gen AI, we can allow buyers to look at different things in different ways. It is helping us with the user aspect of it. Instead of just traditionally giving them reports and telling them what to do, they can now ask questions and come up with ideas.
AI is also helping with efficiency, automation, and challenging us to present data in different ways. One of the best things I like is personalization. Instead of just having a generic model, you can personalize a lot of things. On the other hand, there are challenges like data ethics because we want to make sure there’s no bias. Another thing is data governance becomes a lot more important now because we are allowing the data to be open for anybody to consume. It’s not a report anymore which you can go on. It’s more of people asking us questions, and you need to figure out hallucinations. For this, you need strong data governance. Even though it’s traditional and everyone’s been working on it for a long time, it’s important as we expose the data and make it available.
Beverly: So if I’m hearing you right, some of the skills around yesterday were around different kinds of programming. SQL, Python, PySpark, engineering, computer science, statistics, and those kind of things. But you’re saying today, it’s about different kinds of skills like understanding the ethical components and asking the right questions. Could you expand on that?
Akhil: You still need to have the traditional skills. Since AI has become more pervasive, you need to find the next steps for a data engineer to stay relevant in their role. We’ve been doing MLOps and ML engineering. To me, the next best thing is prompt engineering, and how the technical team can work with the business team to help them learn and ask the right questions. At the end of the day, you have so much data, and depending on how you ask, you can get 10 or 20 different answers. So, the technical team needs to understand how the business works so we can figure out how to work with them. To me, prompt engineering is something which is very important now.
Beverly: Akhil this is an interesting topic because people using prompt engineering today are the coders of yesterday. Five to 10 years in the future, they’re not going to be the coders of yesterday. They’re going to come out with prompt engineering, and that’s where they’re going to hit the ground running. When we ask a question and we get an answer, we’re thinking about a grid and programming and we’re imagining how the question needs to be rephrased and why it didn’t come out as intended. These new talent, they don’t have that. How will that modify the way we operate? If it gets down to the language and focuses on the user and the person, is that going to inherently change the way we view our data?
Akhil: When I look at new talent, I feel they are smarter than us because they are growing up in an environment where AI is available everywhere. Every day, as part of their education, they are using AI and definitely growing the skill of how to ask the right questions. What they need to focus more on is figuring out how to write the SQL and whether AI is giving the right answers. This is where I feel like the senior people come in the picture because the younger talent will know how to use AI in a much better way than the programmers who’ve been doing it for a while. But the senior programmers need to help them make the right choices.
For example, I had a conversation with a programmer who had a billion-row data set, and they needed to iterate to find something through it. So, they used AI to figure out how to do it, and it returned a Python program that processes the data set row by row to do the job. Then we started talking about it, and I told the programmer to use the user-defined function in PySpark to create it and that it’s not an iterative process. This is how we help them learn.
From other evolving roles where there’s a fear of AI replacing roles. I feel that is not true. You need to embrace it because AI could be a good partner for you and will help you grow much faster. If you believe that coding and similar skills are becoming obsolete, it’s not going to help. When AI is combined with your skills, you become more powerful and efficient.
This is true for both generations. A senior programmer needs to embrace AI to help them, and the newer generation that’s already equipped with AI needs to get help from the senior generation to learn the best practices. In my mind, AI is not a way to skip education. But after you learn and go to the industry, it helps you grow faster when compared to someone who started 20 years ago. I had to go through these multiyear projects and different situations where I learned a lot. With AI, this cycle could be shortened. They still need to learn but don’t need to start from scratch.
We know very well where the industry is moving. With AutoML and everything, it’s like you don’t need to even think about which model to build, I’ll let the computer do it for me, and it will tell me which one is the best, and I’ll go from there.
Beverly: I had someone that I was mentoring, and I also serve on the board of the Master of Science in Analytics program at Georgia Tech, and they asked my opinion about joining that program. They were afraid to get a Master of Science in Analytics because it’s going to be outdated because AI will take away all of data science jobs. What do you think about that statement?
Akhil: Right now, there is a lot of space for AI to grow before it can take away data analytics jobs. I don’t think we are at a point where AI understands business processes yet. The intricacies of processes, still, it has some time to grow at that level. You can have models where AI can help you genetically. But taking those models and fitting them to your own needs is something you need to know.
Beverly: Yeah, it can take this one little mechanical piece and maybe do something to improve it or whatever, but it doesn’t see the big picture or what the problem you’re trying to actually solve.
Akhil: Plus, you need to figure out the parameters that apply to a particular situation, which will help you get started, instead of building everything from scratch. But, at the end of the day, there’s still time before it can solve any problem.
Beverly: With the use of AI, we’re seeing more prompt engineering and asking the right questions. You also talked about the ability to recognize some of the inherent biases and dangers from an ethical standpoint and business soft skills. What do we need to do with our talent now that we’re moving in this direction naturally from the evolution of AI? Do we need to re-skill them? Are we going to be upskilling? Will they evolve that way anyway?
Akhil: Right now, we need to consciously upskill them because technology is moving fast, and we are all moving towards the catchy stuff. But we have this existing talent who has spent their careers bringing us to where we are. I think they are ready to upskill. You also need re-skilling, but I think that’s for a different set of talent. When it comes to technical talent, like data engineers and data scientists, they need to upskill and start thinking about gen AI and prompt engineering. There are other talents like data analysts who need to re-skill. Data analysis is a term which we use very openly and easily. In one company, a data analyst could be a data scientist. In some companies, they are just BI developers. Data analysts play an important role because they understand business. So, depending on the persona, they need to be re-skilled or upskilled. For technical people, they often need more guidance. They need a path. They may see that they want to start working with AI and need to know what the next step is doing that.
I read somewhere that AI is a paintbrush where you have to paint your own story and path. If you keep on thinking that AI is going to take your job and you didn’t put in any effort, it’s going to take your job. But if you think about AI as a paintbrush that can help you in learning and in doing what you do today, you are painting your own story. You and AI will basically be working with a different relationship now.
Beverly: Okay, so think of AI as a tool, not the thing in charge.
Akhil: From a coding perspective, AI creates all kinds of productivity gains. I’m concerned that companies might think they no longer need to hire data engineers if AI can code. However, the need for data engineers will still exist, but their roles will evolve, and the need will shift.
I don’t think in the next few years, people will be sitting around doing data ingestion anymore because it’s a repetitive task and AI will be good at it. But now, what to do with that data is something you need to focus on. The focus should also shift to understanding company strategy, data, and how the data can save money or help with efficiency.
Beverly: I was speaking to an information systems class about AI and what it’s doing, and a kid asked me if he should pick a different major because AI is going to be all over this. And I told him, “Of course it’s going to be all over this. But it’s also going to be everywhere else.” Like, I can’t find a place where it’s not going to be. Are we opening the door to where outsiders are taking the traditional data science person’s spot?
Akhil: I’ve seen a lot of people move from other majors to do traditional data science. I think the way they can enter the industry becomes much easier. But, at the end of the day, you still need to work with the business, understand what you’re doing, and focus on how it’s going to help. It’s not going to be about technology, it’s more about how you can provide value. That’s where I see the shift going on.
I’ll give you an example. Five years back, I taught a data fundamentals class at the University of Cincinnati. I was very surprised to see more cybersecurity people reading and learning about data fundamentals than IT majors. I asked why, and they said that cybersecurity is crucial and we need to understand everything happening in the company. Analyzing and gathering data is essential for evolving our cybersecurity approach, making us more proactive rather than reactive. Graduates today understand the importance of data and how to use it effectively. This is a change we’ll keep on seeing every three to four years.
Beverly: If someone asked for advice on what to major in to work in this field, given the current confusion, what would you tell them?
Akhil: I’ll tell them to stick to the basics because that’s a core skill you’ll never lose. For example, data structure is important for anything you build. At the end of the day, you also need to know what AI is telling you and if it makes sense. Even though the cycle is reduced, it’s still there. If you don’t know what’s right or wrong, you could end up harming your company or team. Basically, you could make a choice without understanding whether it’s right or wrong.
Computer science degrees and all that are important. But I think slowly AI is going to start taking over course-wise. I think it’s going to start seeping into core computer science where they’re not learning about the data structures and all that, but they’re also learning about how to use and develop AI. To me, education is still needed, and without having a solid education, it becomes hard.
Beverly: It is interesting because the whole field is changing so much, and it’s going to impact the way we work. When you think about developing data teams, it makes you question if it’s more about language and how we ask questions. Who would have thought 30 years ago that unstructured data and movie-like scenarios would become reality?
Akhil: To me, you need to learn the basics. You don’t need to learn ChatGPT or anything else because those tools will come and go. But you need to understand what’s happening behind it.
Beverly: As far as losing jobs, don’t be constantly thinking that AI is going to take your job. However, someone that embraces AI is going to have an advantage over you.
Akhil: Absolutely. You should always be a lifelong learner because technology is going to keep evolving. A few years back, things used to change every couple of years. Now it’s like every three-to-six months. So you need to learn the fundamentals and need to be comfortable knowing what’s happening because you’re going to get a different flavor of everything.
Beverly: It’s never going to disservice you to go ahead and get your fundamentals down.
Akhil: Absolutely not. I strongly believe that strong fundamentals will help you at every stage.
Beverly: What final piece of advice would you give to somebody who is looking to develop and grow data teams in the world of constantly changing AI?
Akhil: Anybody trying to build a data team needs to check if people have the core skills and not just focus on AI. I work in the finance sector, and Excel is still the most desired tool.
Beverly: Oh, you’re going to say that out loud?
Akhil: Absolutely. For a lot of people working with the users, in the end, you give them dashboards and data they’re downloading in Excel. I think AI will help a lot in that. At the end of the day, these core basic skills are something you should have, and it doesn’t matter how AI is changing our lives and data teams.
Beverly: Thank you so much for talking to us, Akhil Mahajan, Director of Data Analytics at Procter & Gamble, about how AI impacts the way we develop and grow data teams.
Akhil: Thank you for having me.
The impact of AI data teams is both promising and diverse. This insightful conversation between Dr. Beverly Wright and Akhil Mahajan explores how AI can enhance data teams, driving both company growth and individual development. By staying informed about these advancements, data teams can unlock their full potential and make a substantial impact on the organization’s success.
Explore the full catalog of TAG Data Talk conversations here: TAG Data Talk with Dr. Beverly Wright – TAG Online.