#588 Don’t Use AI to Do More. Use It to Solve Bigger Problems with Bala Muthiah

Mehmet: [00:00:00] Hello and welcome back to the opposite of the CTO Show with Mamet today. I'm very pleased joining me from the west coast of the US Bala Muthiah, he's Director of Engineering in Silicon Valley. Uh, we're gonna discuss today, as you might guess, from. Bala's title, director of Engineering. So we're gonna talk about leadership.

We're gonna talk about, you know, the move that Bala did, you know, into leadership role. We're gonna talk about how AI is amplifying also this today and his points of view on this topic. Mentorship, you know, and any, everything and anything you know, related of course, as we discussed always to tech and startups.

Bala, without further ado, thank you very much for joining the show today. I really appreciate that. Just, you know, quickly, maybe a little bit about your background and then you know, we can start the discussion from there. So the floor is yours. 

Bala: Thank you. Thanks a lot for having me. It's really great to be here and I appreciate the time.

Little bit about me, huh? I, as you mentioned, I am here in the [00:01:00] Silicon Valley. I'm working as a director of engineering here at a company, but my roots started back in India, so I started working straight out of college as an engineer and was there for a few years. In 2009 is when I moved to United States to the Bay Area, pretty much in the same area since, at that time, almost, uh, 17 plus years now.

And this is a time where I transitioned also like around 10 years ago from being an engineer to an engineering leader. So right now I'm leading teams and then I'm also working with different startups on helping the founders. This is predominantly pro bono, advising them and also doing what I can to give back because I came here, I am here where I am because of the people who gave to me.

So I wanna do the same. So that's me. 

Mehmet: Great. And, uh, thank you again Bala, for being here with me today, and I appreciate, you know, the time, um, for sharing also your experience that I'm sure [00:02:00] everyone will benefit out of it who is working in tech. So the first thing, uh, you know, I'll ask you like, you moved from outside output focused engineer to people, first leader.

What triggered that shift? Like, what, what, what, you know? Let you to, to, to choose this path, I would say, and how this changed the way you operate day to day. 

Bala: Yeah. The what made this change is a little bit, um, unconventional, I would say. Uh, so this was somewhere around 10 plus years ago. So I went to a school, this is all, I was in an as an engineer, so as part of the corporate social responsibility initiatives at work, this was my previous job.

We went to a school, like a public school in us. So for context for international audience. So public school in US is. Funded by the government, but the range, the quality of the education depends upon which neighborhood your school is. If you're in a well invested neighborhood, the public school will be really good, right?

If you're [00:03:00] in a neighborhood which is under invested, like the population economies, like the, uh, income is lower, it's not gonna be heavily invested because most of these are driven by tax money. So if you're in a good neighborhood, which has. Expensive houses, higher taxes, better schools. You are in a neighborhood which is not very invested in lower taxes.

Low funding goes to public schools. So I went to one of these public school schools to volunteer, right? This is to go paint the schools, like usually companies go and do this. When I went there, I had a chance to talk to the principal and they talked about I would love to have someone come and teach computer science as a one-off thing to the kids.

This is an elementary school, like grade three, four, and five. So I started doing that, like I do Wednesday and Friday mornings from seven 30, from eight to nine 30 ish, like just an hour and a half, three classes. So this was, this is, I would say, the moment that changed my life. This is 2015 ish time. [00:04:00] Seeing those kids grow, seeing those kids learn and then develop, that really gave me the sense, Hmm, I wanna do this for a living.

And being in tech, being an engineer at that point, gave me that transition. Oh, being a manager is very similar. You invest in people, you grow people, and there is a great sense of satisfaction and gratification. So that was my pivot and switch to leadership. 

Mehmet: Nice story and, uh, very touching, I would say about that Now, um.

We know that sometimes, uh, we have strong engineers. Like when we say strong engineers, they are like, really, they know what they're doing. They're very well versed. Um, and, you know, part of any career path, like for anyone in any domain, they aim for becoming leaders, right? And, and of course, building the next generation of people who can be doing the same thing, but.

We hear it sometime, it's [00:05:00] not a path for everyone. And sometime people insist of going to that path and we see them kind of doing some mistakes, I would say. So from your experience, and because you've, you've watched this space for a long time, what are like some of the biggest mistakes you saw engineers make when they step into leadership roles?

Bala: Oh, this is very, very true, right? Like you mentioned about, um, strong engineers. I think this pretty much anyone who is going through. Are gonna have, right? Like I wouldn't claim myself as a strong engineer back in the days, but when I transitioned, I made those mistakes too. Like even anybody, like strong or not, this is gonna be a common, uh.

Pattern because this is part of learning. There is no school. You go to become a leader, right? Because you see, you do it in the job, you get hands-on practice and then you switch. So one of the things which I have done as a, um, mistake or like something that, oh, probably not a right way to do things, which I see some people tend to do a lot [00:06:00] today, is when you come from an engineering background, you are.

Always about problem solving, right? When you see something is wrong, when you see something is not going in the right direction, you have the urge to fold your sleeves and step in and solve those problems. That's what an engineer is. Debug, break solve problems, and when you become a manager, when your team is going through a problem.

The first instinct is to jump into it and roll your sleeves and help them fix. So this is something we call as like a, a rescue thing, right? Like you jump in, you want to go rescue because you, you see your team is, uh, now going through a tough time. This was one of the things I did early, early in the days, and this is actually an anti-patent because you are.

In a way robbing the opportunity from your team to learn, you are getting the things that they would've done otherwise. Yes, you stepping in will definitely help. You can accelerate the solution however you are taking the [00:07:00] opportunity from them to learn. So that was one of the biggest things I do. And similarly like then you learn couple times and then you tweak.

But the biggest one I would say is getting into the rescue mode and jumping in. Too early before your team actually needs you. 

Mehmet: Great. Now things are changing fast, but, uh, as we know because of ai, right? So it's, it's disrupting, um, you know, verticals. It's disrupting even the way we, we do our jobs when it comes to leadership, especially in engineering, of course.

Um. When I was preparing, I saw like you, you, you mentioned you frame AI as leadership amplifier, not just as an automation tool in real life, in a real engineering organization. How would that look like? On a high level, and then we can maybe dive into some of the details. 

Bala: Yeah, totally, totally. I, I've seen this, of course, AI, as you mentioned, is [00:08:00] rapidly evolving, right?

I, I jokingly say that whenever we go on shows or talks, by the time you start the show and end the show, already, something else would've happened in the world of ai, right? It's happening so fast. So why I say this as an amplifier, or I see. This is because things are happening so fast, we are able to solve problems so fast.

So that means there's a good positive thing, which is, hey, I'm able to get to a problem faster than I could ever before. Also, it's gonna amplify. And all the flaws and gaps in your system, right? Think of it like a lens, right? Like a zoom lens. It is gonna show everything big. Like you can do things faster, you can do bigger things, but it's also going to open up and show your F floss.

If you have a small crack, then you accelerate. It's gonna be become a big crack. So this is where leaders have to be very, very intentional and very. Deeply, uh, curious to understand what are the existing gaps in the [00:09:00] system, flaws in the system, the cracks that when you start building an accelerating is gonna break.

So it is a great amplifier because you can do a lot more things very, very quickly. That comes with the cost of those inefficiencies that you have. The dysfunctions you have is also gonna accelerate, meaning it's gonna also break faster. So this is where you as a leader make a huge difference 

Mehmet: right now.

Let's break it down a little bit more to the audience, Bala. So we know, you know, when we talk about AI and, you know, the way it can empower us, as, you know, knowledge workers as we they call us, right? Mm-hmm. So, uh, you know, automating tasks and sometimes, you know, even getting some insights. So where have you seen AI actually improve leadership decisions?

Rather than just productivity, because productivity, I think, you know, everyone knows now that AI can create your Excel sheets. It can go and, uh, do your repetitive task. It can do [00:10:00] triage for your emails in the morning. It gives you the, so, so we know all this, but when it comes to decision, because you know, majority of the time when I ask, you know, whether.

VP of engineering or CTOs or you know, especially in tech. So this is majority of the time they tell me they spend it in, in decision making, not in actually like productivity kind of thing. Because this is for the team, it's not for them. So can you give us examples like where AI actually can improve the decisions for, for leaders in, in technology?

Bala: Yeah, no, totally like decision making is. Probably the most important thing a leader brings to the table and how good your decisions are is gonna change the fate of the company, of your team, of your organization, anything. So with ai there are, there's one thing that has made the world, I would say, easier in way two, which is access to data.

Right? Before, when I say data, I'm not [00:11:00] talking about raw information. I'm talking about. You are a company, you are trying to build something, sending it to your customer. You can send it like do it so quickly in a very rapid fashion that you can immediately start getting customer feedback. Back in the day, what would take, let's say a month or two months to ship a product to a customer for them to get used to it and then give feedback, and then you use that to improvise how you.

Develop or improve your product. That cycle has, that cycle has shortened. So you can quickly experiment something rapidly, go find out what's happening, and then you can take decisions. So for a leader, this is, I would say, one of the best times to be because. You can know what someone is thinking about their product very, very soon, which was not the case back in the days.

So that is a differentiator. So when you have access to such a quick, short feedback loop, what can you do as a leader to improve your product? Going to be a big differentiator. So access to data before. If you as a leader [00:12:00] wanna know how your business health is performing, wanna know something specific, you have to ask your team.

And they have to go build dashboards and they have to get it all. Now, it all. Readily accessible to you in a chat interface, right? You can just type in a question and it's gonna get you the data you want. Then you can quickly act, you can quickly, uh, perform what you have to perform that is important to the business.

So access to data are the speed at which you can get insights. That's the differentiator for decision making in the new world, 

Mehmet: right? So part of the decisions, if we, if we want a little bit kind of, uh, go more deep, is. Dealing with conflicts, right? Yes. And maybe sometimes, uh, uh, you know, searching for growth opportunities.

Right. So is there also a practicality for AI to help leaders in, in these two aspects as well? Because I know you have a point of view on that. Bah, 

Bala: yes, yes, definitely. Actually, this [00:13:00] is more, again, um, we are in the. Early phases of ai. So we are learning a lot about ai, right? Nobody can really predict fully, oh, what is, and how it is gonna happen.

The best thing to do is just be curious and try it out. So the opportunities, right. Before we, we talked about productivity a little bit. You mentioned about productivity. That's like the foundation. Most of the companies are talking about how it'll improve proactivity. So what happens? Gaining productivity is not the biggest thing.

Right. It's what comes after the productivity. For example, let's say a task is taking 10 hours, right? Or for a team, let's say taking 10 weeks to build something. Now with ai, let's say you can build it in three weeks, right? So that's, you can do it in 30% of the time compared to before. So now one of the common pitfall is, hey, now I have seven more weeks.

I'm gonna go do three more things, right? [00:14:00] Or, oh, I'm gonna sit and do nothing. That's a another anti-patent. Or, oh, I can now do it in three weeks. I don't need that much. Uh, that many people, I don't need that big a team, so I'm just gonna shrink. So those are all antipas. What you should do is go after bigger problems.

We are always constrained by our resources, right? The problems we go after, like even a small company, big company, businesses, anything you go based on the resources you have now that has expanded. That has grown. So for a leader now, growth is more expansive. I use the word expansive instead of growth necessarily, because when we say growth, we always think about vertical, right?

Mm-hmm. We think growth is like increasing something that you have expansive is like going wide. So that is an ability that AI is gonna give to leaders and it's up to a leader to use it and choose it to use it wisely, right? The expansiveness that is gonna be the accelerator and that will make or break a business in [00:15:00] the new world.

Mehmet: Absolutely, I agree with you. Like, uh, this is what it'll make it or break it. Now there is nothing. But as we know, that comes without risks. Right? So what are like some of the risks of adopting actually the AI without, you know, maybe looking back at there, maybe leadership style. And I'm asking you this question because you know, like this is the same thing when.

And I'm sure you, you saw the hype in, in Silicon Valley and people saw it like everywhere else. Uh, like everyone want to adopt AI regardless the function. And we find out that actually you need first to do some pre-work before you can make the AI work for you. So when we take, and this is just, you know, having your processes like documented, having your data ready and all this when it comes to leadership.

Like what are like the prerequisites I would say that a leader has to have before they can adopt AI without having the risks that comes with the ai? 

Bala: Yes, [00:16:00] true. Like these transitions have happened a lot before, but this time it's much more. I would say, uh, exponential, like things are going really fast. So the first thing to do, we talked about earlier, right?

It's we have to unlearn some of the practices because AI is gonna amplify everything. So if you have a productive system, it let us say it's gonna make it 10 x. So if you have a flawed system, it is also going to make it 10 x. Everything is gonna amplify. So back in the time, let's say you have a certain.

Behavior as a leader. Let's say someone is micromanaging, right? Mm-hmm. That's a common pattern. Leaders fall into. 

Mehmet: Right 

Bala: With AI that's gonna amplify, because now you are gonna be more invasive. You are gonna have access to data a lot. So you tend to lean more into micromanagement because now you can look at every nuance of what's happening in a team, and you would have the urge to go solve it out of goodwill, right?

You wanna go solve. So that's an anti-pattern to think about. [00:17:00] So if you don't drop those, this is. Purely from a leadership perspective, if you don't go do that, if you are carrying your whole old self to the new world, it'll break. So drop the. Things that was a minor thing before was not good thing as a leader, remember, that is gonna get amplified.

So be intentionally, uh, reflecting on you do a retrospective before you start this new leadership journey with AI and then drop those pieces. It's not gonna help. It's gonna hurt more. Back in the day, maybe it was okay to have. That slight, uh, unaccepted behavior because you are good at other things, but now that's gonna amplify, so it might completely change the formula of the equation.

Mehmet: Absolutely. You know, because the ai, it, uh, it's like a very good imitator of what we do. So, and, and that on steroids, like if we have micromanagers, we become micromanagers on steroids and team would hate it. Of course. A [00:18:00] hundred percent. Like it's a nice example you gave. Now, let's say when, when you dealing with, with.

You a team, I would say like, um, and of course they, they would be empowered also to use ai, right? 

Bala: Yeah. 

Mehmet: Now, what's your, what's your take Bala on also encouraging a team to judge on the AI outputs. And I'm asking this, whether it's like code generated by ai, should they take it or they should, like, again, think about it and you know, okay, let's say, let's have a real review on, on that.

Or maybe sometime. Uh, and, and we've seen like some example that came out in the news where, you know, people completely gave it the ai Yes. They dropped the ball kind of. So, and I've, so I've also spoken with people like, who are also like. Head of engineering CTOs and they tell me like, you know, they [00:19:00] have, you know, these kinds of human guardrails when coming to take decisions.

So we, we don't a hundred percent now. Where does the human judgment still, in your opinion, outperform AI today? 

Bala: Yeah. This is gonna be the critical differentiator, right? Between winners and who are not making it. So human in the loop or human decision making is still gonna be critical, right? Yes. The rate at which we are gonna do is changing.

Rapidly, all those things, but having the human is gonna be key. I'll tell you, uh, this where if you first, as, as a leader working with the team to figure out what are the tasks that are okay to. L use AI for? What are the tasks that would definitely not, not use AI for? Right? You need to have this principle or a protocol as a company.

Depends upon the business, depends upon the culture, depends upon what kind of data [00:20:00] you're dealing with. You need to have a playbook. It's almost like a compliance playbook where c, certain areas you do not. Let AI at all certain areas, you let ai, but with full control and certain areas, you can let it go on its own.

For example, documentation, internal documentation, completely let it go. It's okay if it messes up. You can do. Dealing with data, dealing with access that are critical and sensitive, and privacy identity things do not let it near if you let it near have very, very strict, tight guardrails. Right? So those principles, based on the business one can come up with.

And another thing which often people don't think about, this has nothing to do with data or a company. It's with people, which is, I think of our work we do in, we can classify that into two, but, uh, like. Two accesses. Like if you take the any tasks we do, you can think about like easy and hard, and then you [00:21:00] can think about, um, fun and boring, right?

If you put into these four uh, categories, the easy and boring things. That, that give, give it to, uh, ai. Like it's easy, it's boring. What do you wanna do? AI can do it pretty well, right? Documentation, gathering information, summarizing, stuff like that. And you can go into all the four different types. The one, the hard and fun part, I would say keep that.

Do not give that to AI because that's where your team is gonna thrive, learn and grow Your team is there, an engineer is there, or any role. They wanna solve fun problems, so they wanna solve hard problems. So the hard and fun ones, save it for them. Just because AI can do, just don't give it. This is nothing to do with compliance or anything, but this is how you make your team excited.

Make your team happy. And stay relevant and, uh, continue as an expert in your team because that's gonna be needed. We talk about our own agents fully running company, but that's in the future. We don't even know where they are. Of [00:22:00] course, models are getting better. We are getting higher quality input, but we are not there yet.

Don't let it run on its own, like have a proper. It's evolving. Yeah, it is evolving. It's coming from like I would say, an infant to a toddler. Maybe now it's getting into a teenager phase. So as a parent, if you think, kinda understand how it is and how much control you would give, so assume the same and give that autonomy, but have some control.

Mehmet: Yeah, of course the controls has to stay there. Um, for sure. And yeah, it's good. You mentioned these, uh, high sensitive examples, like, for example, anything with regulations and you know, so, so where, where you keep in control now as a leader, but of course part of the things that you do is. Mentoring, you know, um, your, your team and maybe even maybe you volunteered to, to mentor people outside as well, uh, in, in, uh, in, in [00:23:00] real life.

So you mentioned, of course, I'm going to my notes, uh, that before, you know, I was preparing, so I saw, like you said, AI is democratizing mentorship, right? Um, so is that where anyone can, you know, because just I can go to chat GPT or Claude or Gemini or whatever model and they can chat with it, or is it because, you know, I can.

Of course, it's you who cannot do it. Or Bala can go and put his knowledge, his, his experience in kind of what people call it, uh, uh, MD file skills. And then you create, you create an avatar of Bala. And then at any time me, I can go and talk to, to a a, uh, to a, uh, digital version maybe of Bala or, or how does it look in, in, in real life?

I'm asking you this, uh, by, because we know still till now, of course, as you said, like we are in the teenage [00:24:00] phase maybe now of the ai. Mm-hmm. We know like that we have a lot of issues with AI outcomes or, or outputs. I would say, for example, you know, bias, uh, sometimes it's still hallucination. Uh, we didn't solve, we, we, we crossed like long way.

I'm, I'm, I'm sure of that. I can see it. But still, you know, we have these, um. Kind of hiccups here and there. So when it comes to, to mentorship, uh, how does it look like with the current status? Of course, I don't know if by, you know, the time will is the episode, which is usually after two weeks from, from, uh, recording.

And for the transparency with my audit, I tell them usually when I record it. So we are recording on the 24th of, of March. So I don't know if by 1st of April we're gonna have something major, but with the current, uh, with as, as per the time of this recording, Bella. How the, you know, your point of view of the ai, um, you know, democratizing the mentorship looks like in 

Bala: practice.

Mehmet: Yes. 

Bala: No, this is [00:25:00] something I've been, um, like you said, like mentorship is a big part of my life personally. Um, as I mentioned earlier. I am where I am because of my mentors, right? I am standing on their shoulders now and I see AI is really enabling everybody now in a better way. So you asked two options, right?

Like the first one is more, oh, is it something that a person can now interact with an LLM? And get mentored, right? That's one. Or are we creating a model or of our own agent or our own clone that somebody could go and be a mentee and that could mentor? I see the third one, which is both of these humans, right?

The mentor and the mentee are still gonna be in the equation. It is just AI is gonna help them assist in this conversation better. Because today a mentorship is very one-on-one and very relational, right? You don't want to. Make one part of that synthetic. If I'm talking to an [00:26:00] LLM mentor, it's gonna be synthetic.

Yeah, it's gonna gimme all the pleasing answers I want. It's gonna say my ideas are great, that I'm doing phenomenal, but it is gonna get me into delusion. Right. Only a real human can understand and unpack. What's happening because there's emotions involved in mentoring. Nobody goes to a mentor purely from a skill-based thing.

If you're going, then you can always Google that specific skill, or you can read a book, you go to a mentor for real life experience that cannot be synthetically created. So what I uh, mean by AI is democratizing is. Now instead of two people in a mentor mentoring relationship, you are adding a third per third entity, which is ai.

So that is going to actually monitor, observe how the conversations are. It's going to help both the mentor and the mentee to be a better version of themselves. So I've seen this when I mentor people, like some of the questions that I get asked. And then I respond right on the fly. It's not that I have time to [00:27:00] go prepare, and that's why mentors, as you get more experience, you are gonna be more valuable to your mentee because you might have seen things, but there are things I might not know.

This is where, based on the conversations and based on how I. The questions are like, it's gonna help and assist me. Same way with the mentee, right? They can get different perspectives on top of what I'm providing. So I would say this is like a coach helping us, both mentor and a mentee. Accelerating both of us for the better, and then eventually it'll build more context, and then it'll make our whole journey more successful.

Today, a mentor and mentee, there is no third person, so it's hard to know if it's going well or not. It's hard to raise the bar, right, because it becomes personal. This is where these tools will help. And of course, opening into access, right? What do you have to now wait for an expert? Is no longer needed.

You can quickly check with something. Those little things, like you said, the option, like the first, uh, version where you can still engage, [00:28:00] but I wouldn't relate as a mentor, right? So that's gonna be a tricky one. Maybe, maybe another part we didn't talk is maybe you can trick AI as a mentee and you can start teaching.

Maybe you can get better at that. You can like kind of do a mock mentoring session and ask it to act really, uh, naive and ask questions or you really. Hone your skills, but I see there's a third person in the room who's gonna add value to the both, um, to both individuals. 

Mehmet: One thing I noticed, you know, just as a quick follow up on this is because you said like, we can teach ai, and I think what AI is becoming good at is training, actually.

Um, because when I'm asking, for example, to explain to me, and you know, this famous prompt, like explain to me as if I am 

Bala: mm-hmm. 

Mehmet: You know. Beginner or explain in layman terms or whatever, you know, the models are coming really, really good in this. Yes. And you know what I'm thinking? So maybe you know, the onboarding process, and I'm sure at at engineering, you know, departments the same thing [00:29:00] happened.

So maybe, you know, the AI can take part of this. As, you know, the tone and voice quote unquote of Bala, let's say this is your, your your way and then, you know, give it to them. So probably, you know, there are like, um, you know, kind of things that you can, as you mentioned, like it's not like a fully a hundred percent.

Yeah. You go, you give it your character and mm-hmm. It does the things by itself, but it comes kind of a supervised, um. Journey also as well too. So you give feedback and then you teach it to become a better mentor sex time. So, you know, it's a little bit philosophical. I know, but I, this is my own point of view on this.

Bala: True, true. No, very true. People are getting very creative about how they use, right? They're running simulations. Oh, this is how famously now we are seeing in the news that, oh, mark Zuckerberg is creating his own, uh, clone or like a a IC or to help him with day-to-day. So people are running predictions, right?

Like, oh, this is how I'm thinking. What will. What will my leader respond to? Right? This is a document [00:30:00] I would like my leader's perspective. These are their historic previous comments. So it's gonna get you enough comments that you can pre-addressed before you actually go to the leader, right? Right. So I think this is like a step N minus one before you go to the final step.

So you'll go with much more better version to the final approver or reviewer. 

Mehmet: Right. Vale. I, we, we didn't discuss much and I don't have, you know, a lot of questions there, but just out of curiosity, when it comes to day-to-day operations, so one of the thing we hear about, especially in, in engineering departments is, you know, the high performance but sometimes would come out with the burnout, the fatigue.

Bala: Yes. 

Mehmet: And, and everything in between. So, uh, and you know, of course this is something. There is no escape, especially if you are in, in, in, if you are working with a company in a, uh, I would call it in, uh, um, high growth, uh, phase where everything needs to, you know, be finished fast and you have like very strict deadlines and usually, you know, [00:31:00] engineering in general, you know, and, and, uh, you know, being in, uh, this environment is not an easy thing.

Now, when it comes to ai, is it, are you seeing like it's helping also? To keep this high cadence of performance lowering burnout, or actually it's making it worse. Like what's your view on this? 

Bala: Yeah, this is, again, still early. Like it depends upon how it is being used by the people, right? So it's gonna come to the person who's deploying these and using these techniques or tools, and in a way, I don't see this.

Actually helping in that kind of way, right? Meaning, because we are all excited, we all want, there are two types of burnout, right? One is, hey, I can do a task faster now I'm gonna go do more tasks, so that's gonna burn me out, 

Mehmet: right? 

Bala: So that's one of the things, right? But even before even we go there, the phase where most of the most [00:32:00] of us are in today is there is so much happening in the world of ai, in the world of technology.

We all try to stay on top of it. We wanna go and know everything new that is coming up, or there's a new model, I wanna experiment it, there's a new app, I wanna experiment it, or there's a new, um, paper that has come out. I'm gonna go read that or there's a new skill. I wanna go get that. And this is actually causing a lot of burnout because you are constantly in this chase where, oh, I need to exhaustively know everything.

That's never going to be the case. I have done this before too, like couple. Months ago, I was in this cycle where I was, oh, I wanna be on top of everything. I wanna know everything. I'm gonna try everything. And then I burnt out. I gave up because it's never gonna be possible at the rate it's come coming out.

It's like drinking out of fire hose, right? You cannot do that. So first it stop. That itself will give you a mental clarity, right? So that's the first burnout I wanna. Caution people, which is very evident in people I talk to. [00:33:00] Either they're burnt out because of, oh my God, so much is happening. It's overwhelming.

Or they're burnt out because, oh, I'm giving everything I can, but it's still. I feel irrelevant. I, I feel I'm not up to the game. I feel everyone is ahead of me because all you see in the news and media and social network and like things is, oh, I built an app in one day. I built an Yeah app in half an hour.

So you start questioning and you burn yourself out. The other burnout you talked about, oh, I can do a task in two hours now, which was taking 10 hours. Some teams go into the zone of I'm gonna take five tasks and do so. That is a burnout because the decision making has not changed. Right. You are still in the loop.

Yes. A project. Only the execution part, the rate at which you can code that has shrunk, nothing else. You talking with somebody in a different team, validating something, getting requirements, getting to a customer, those things are not changing. So they are still there. So if you try to do five [00:34:00] projects.

It's gonna burn you out. The best thing is to go after bigger problems. So your lifecycle doesn't change. You are still doing the same, but you are just going deeper. A simple example is with ai, you can either go, I would say like a mile wide and inch deep, or you can go inch wide and mile deep. So go inch wide, mile deep, so go deeper in a problem so you solve bigger problems.

Not more number of problems. That is a way to avoid burnout. Otherwise, we will be like, I have personally felt this so many days. Like I wouldn't say I have found a way to come out of it, but I'm slowly getting out of that burnout zone, acknowledging that, oh, I cannot stay on top of everything all the single time.

And yes, I'm gonna feel left out. I'm gonna feel really dumb because new things are coming up like that is the learning exercise, right. 

Mehmet: A hundred percent. And you know, of [00:35:00] course, like, you know, I have technical background. So by, um, by default, you know, my brain, whenever I see something new, I would say, yeah, I want to go and try this out.

Um, because, you know, of course I want also to be on top of everything. Whether, if I am, you know, doing my outside of the things or whether mm-hmm. I'm preparing for, for, to talk to someone on the podcast. So I, I, I need to be, you know, at least up to date on everything and anything now. To point I, maybe this is my.

All experience on this. I had like similar feelings, which you just described now. Oh, like how come these guys, you know, all over the place, they're doing these fantastic things. Like as you said, like you have the self-doubt. Am I doing something wrong? Am am I, you know, not following up properly. Then after a while I figured out two things, which helped me a lot.

And of course yours is the best, you know, to go about the important stuff, go deep. Now the other thing which I started to see is there is a lot of hype, especially if people who hangs around, you know, the [00:36:00] socials link it in more ex. You know, and you see these, uh, you know, there's someone who calls, not my term, but someone who calls them the bros.

You know, like these guys who, you know, like the, the, the day something comes up after six hours, they recorded a full YouTube video Yes. Showing the people How so, of course, like, you know. And I think this is part of the ex experiments, uh, and experience in the same time experimenting and experiencing, you know, this.

So, and you figure out, okay, I know why these guys are doing this. They're just looking for visibility and you know, I'm not looking for visibility. I, because going back to your deep. A thing I want to go do what is important for me. I don't need visibility. The second thing you know with time, you know, you start to learn, okay, how to filter these things out.

Okay. Now, when we use the new models used to come, you know, everyone was used to get excited, but now, okay, the first thing, personally, I do. I go read the release notes. Okay. What they are saying, like what they have improved? [00:37:00] 

Bala: Hmm. 

Mehmet: Does this help me in my day to day thing? 

Bala: Yeah. 

Mehmet: Uh, I'm, I'm not a coder, so if I see, like for example, I'm just giving it, oh.

Like, for example, open AI Codex is now better than, you know, cloud Opus. 

Oh, 

Bala: mm-hmm. 

Mehmet: In coding, I'm. Just give an example for me, I don't code, so I don't have to have the feeling of fear of missing out, right? True. But if, for example, let's say I work on preparing marketing materials and they tell me, Hey, like the new cloud model is, or Gemini model is much better in this.

Okay, let me go and check it out. So I start to mentally, you know, train myself to filter these things out so I don't have this feeling. Of burnout, as you mentioned, uh, feeling like you start to, you know, hit yourself, oh, I'm not following enough, so I, I get like this. And again, it's all. Turns to the point you just mentioned, which I think it's very, very important, is going deep and search for what is [00:38:00] meaningful and bigger problem, rather than just going after the, you know, I would call them the doom and gloom of what people chose us, uh, here and there.

Which I think it's okay because maybe I, I was very young, but maybe when the internet came out, things were like this at that time, you know, maybe. Yeah. But of course not at that large scale. There was no social, uh, media and all this. 

Bala: Yeah. 

Mehmet: But, but yeah. But, but I think your point is very, very, you know, I'm repeating it.

Go deep and solve bigger problems. I think, you know, this is the best takeaway, at least for me. Yeah. You know, uh, one of the best, uh, of the takeaways today, now as we are coming to, to the end, and, you know, I want to close with this topic, part of the. Any organization, an engineering organization is of course not different.

It's all about not only building teams and building systems, but also to establish culture right now, and I know you have your own view about building systems and culture that [00:39:00] adapt instead of age. Now, how does an aging engineering organization looks in practice in your opinion? 

Bala: Yes. I think this is where, um, like for any system or a team that has been there for a while, or like legacy we call it, right?

What we don't have is the learnings are not updated because things are maybe outdated. Like now with AI and stuff, you get the learnings faster, like you can get those things exposed faster. So that's a great, great tool to use it for, right? Hey, find out what kind of debt you have. Kind find out what kinda antipas you have, which you have always lived with, but never had a chance to go visit.

Because yes, it's important, but it's not urgent. So it's gonna be in the back burner. Right? So now they can come to the front burner. So use AI for those kind of taking away debt and stuff. And you mentioned about culture, right? That has always been [00:40:00] very, very important. Part, especially now with ai, your competition can build what you have overnight.

Imagine you are building a great feature, right? Really phenomenal, wonderful insight. It goes out to the market, everybody gets, wow, this is great. And then day after tomorrow, your competition builds the same thing because all it takes is one day now with the all the tools to build something like that.

Mehmet: True. 

Bala: What your competition cannot take away or build is the culture. Right. What can your team do? That's where leaders and going to be more, the leaders who are winners are the people who are not just looking at work, but who are looking at the people who can do the work and building an environment. For those people to thrive, like your competition cannot take away that culture of your team going after the next one going deeper.

There is famous Peter Docker thing of, uh, culture eats, uh, strategy for [00:41:00] breakfast. True. So culture is going to be the critical thing. That's the one. AI cannot come and replicate or synthetically create. Everything else could be duplicated like tomorrow morning. So keep that in mind as a leader and think about what is your team going through, like building the psychological safety in the world of AI is more important.

Building the trust not only in the AI and systems and data, but also within the team. Like build your team, like get people who are problem solvers. 

Mehmet: This is very true and you know, this is what I'm discussing with a lot of people today because, you know, especially when it comes to the world of startups and even like, you know, established companies, as you said, like you, you can copy a feature in no time today, and we talk about, of course we can discuss and we can argue about what is the mode today.

Is it, yeah, you know, the data or, but you know, one of the moats, which for sure it's one of the most important moats, you know, it's the people. People means culture and people means you know, the team [00:42:00] and how you deal with the team. And if you have a fantastic team, you know, I believe, you know, and if you have established a fantastic culture, like you can go after any competitor because you know, like as you said, copying features.

Actually it wasn't, I'm not saying it wasn't, uh, e easy, it, it was difficult, but I mean, it used to take time, but now this is get accelerated, but you can't copy a team spirit, I would say. Yes, a hundred percent. You know? Yes. So, so I would agree on, on this valor, but I really enjoyed the, you know, the discussion today.

Final question, like if people want to, to, to get in touch, do you have a website, any, any need where people can, can know more about, uh, you? 

Bala: Yeah, totally, definitely. They can find me on LinkedIn or my website, bala.com. So, um, I have contact there and reach out to me on email. Those details are there. Always happy to have conversations because the more we need to do now is be curious and understand where everybody's.[00:43:00] 

Like, that's the thing that it is gonna be hard to get. The real value is in talking to people and understanding where they are and what they're trying to do. So always open to those conversations. 

Mehmet: Absolutely. Uh, Bella, I really thank you very much for, for this discussion today. The link would be, or the links would be in the show notes for people who are listening on their favorite podcasting app.

They will find it there. If you're watching this on YouTube, you'll find it on in the description. Uh, and, um. Again, thank you very, very much. I really enjoyed the discussion because I think this is, um, important to understand how engineering leadership and leadership in general in the age of AI should be.

Practiced, I would say, uh, and again, I'm, I'm highlighting my takeaway, my personal takeaway from today is about going deep and solve bigger problem. Um, it's, it's, it's like my favorite takeaway from, from the whole discussion today because I think this is what it matters after all. And of [00:44:00] course the culture and the team and, and you know, being able to leverage the AI not just as productivity tool, but more as you know.

A framework to shape our leadership style. So, so a hundred percent on that. Also, Bala, um, and this is how I end my episodes. This is for the, uh, audience. If you just discovered this podcast by luck, thank you for passing by, and please give me a small favor, share it with your friends and colleagues. We're trying to do a, you know, more awareness.

We're trying to share knowledge with as much people as possible. And if you are, of the people who keeps coming again and again, the loyal fans. Thank you very much and I can't thank you enough because. Since last year, 2025, we started to see that the podcast never, ever passed a week without being in one of these countries.

Apple Top 200 podcast charts in the top list. This is cannot happen by itself. I don't push people to listen to the podcast. So this is all because you are letting other people know about the podcast. So thank you very, very much, and as I say, always stay tuned for any [00:45:00] episode very soon. Thank you. Bye bye.