Great. Okay. So hi, everyone, and thank you for joining. I'm Topaz Gillard, and today, we're going to talk about AI.I know that's a huge surprise, but I promise this won't be your typical AI talk because today I wish to share a fresh perspective on leadership through AI inspired strategies. What if we could, draw inspiration from practice that are used to train AI models but to help our teams grow? And, no, we're not turning your people into robots, but what if we could borrow a few techniques from the world of AI development in order to enhance the way that we lead and support our teams? Let's explore how these principles, same principles behind training AI models can actually inspire us to take smarter, more intentional control over our careers. So what led me to talk about this topic? So let me briefly introduce myself.

I am VP R and D at Voyager two One Audity, again, Topaz Gillard, and I'm also a mother of three, ages eight year old, five and a half, and I also have a tiny one year old. And in addition, I mentor and lecture on career development and leadership. My background is in machine learning computer vision, algorithm development, and over the years, my career has spanned in diverse domains from space industry to biomedical imaging, semiconductors, sport tech, and today in the beauty and wellness industry, and I've worked in organizations of all kinds from large corporates and fast paced start ups.

And as I transitioned into leadership roles, I actually realized that the real challenge isn't just in building great products, it's in building great teams, effective processes, and strong cross functional collaboration. And interestingly, I found that those same techniques that I use to train AI models were actually helping me shape my management thinking. And in this talk, I'll try to bridge these two worlds, the world of algorithms and the world of leadership, and just share practical tools and perspective that help me lead more effectively. So let's first set the stage. Behind every great AI model lies a continuous process of learning, adapting, optimizing, and that's also basically the foundation of effective leadership. And if you like, we can think of building and managing our teams as an algorithm, a set of iterative steps where feedback, pattern recognition, and learning drive continuous improvement. And today, we will explore how we can leverage these same techniques inspired by AI development processes, not only to build excellent AI models, but also to accelerate the growth and effectiveness of our careers and our teams.

At the core of every great AI model lies the data, the data that used to train it. As managers, we we sit on a goldmine of information, but we don't always know how to leverage it, like, respectively. Think of every feedback conversation, every code review, every suggestion from a colleague as raw data, every team retrospective one on one meeting team discussion as a data point in that data cloud that we want to use wisely to improve ourselves, our teams, our work. And when collecting data in order to train a model, two very important questions just come up. First, how diverse and high quality is the data? And secondly, how well have we used it in the training process? So let's examine those two. First, what does it mean to use diverse data as managers? It means gathering our inputs from different sources, people in various roles with perspectives that could be very different from our own. And what about data quality?

So I will share a personal perspective here. The manager that I was, well, not too long ago, that manager thought that feedback happened naturally day to day, and that structured feedback, well, I felt like it wasn't really valuable, like it felt forced. But if I would have looked at it through the lens of a data scientist, I would have realized that it's the same as expecting to train a model with unclear, inconsistent data, and of course that leads to a very different results than when training with very clear, non ambiguous data, and that is why it's so important to both give and receive feedback in a thoughtful and structured way.

And if you are worried that your team's retrospectives could be just take too much time or they will be unproductive, then you can always try some focus questions ahead of time on what's really, really important for you to discuss and just remind the team to think about those key points before the meeting.

But as always, save some space and and time for new topics to come up during this conversation. So we talked about diversity and quality of our data, but sorry, what does it actually mean to make the most of the data that we have? It's not enough to simply collect endless amounts of it. Even in AI, training a model requires processing of the data, not just storing it. And as leaders, we need to translate that information into action. That means closing the loop on feedback and letting data inform how we operate. Well, we mentioned retrospective, so let's take that as an example. Personally, I make sure that each session ends with a clear summary and a review of what has changed since the previous retrospective And of of course, a list of new action items, and those are not just items on a list.

Those are links to actual Jira tickets that we can estimate, prioritize, and incorporate into the outcome sprints because the goal is not to drift from one sprint to the other on an autopilot. No. It's to actually break through barriers. And same but another example is one on one periodic reviews, those feedback conversation. It's far more effective to revisit points from the previous meeting and reflect on what has changed. Because when feedback drives visible change, we foster teams that are engaged, accountable, committed, teams that move forward with purpose and grow together. So what do AI models actually do with all that data? They identify patterns. That's the foundation of modeling, spotting trends in order to create a probabilistic modeling that then allows us to make predictions. What if we applied same mindset as tech leaders?

What if we trained ourselves to recognize signals, early indicators of challenges or opportunities so we could take action and be proactive? For example, can we, spot early signs of burnout or or declining motivation in our team? Can we identify recurrent blockers that surface every sprint, or, can we identify situations at work that tend to trigger resistance for us or our team members? How can we be proactive to prevent friction there? And even when managing upwards, can we identify some kind of a presentation style that makes our manager more receptive to new ideas? And we can even ask, where do cross team tension typically arise? Is it joint deployments, integrations, and and how can we mitigate those early on?

The more we intentionally look for these patterns, the more influence we gain. We become better at navigating complexity, anticipating outcomes, and steering towards the results that we want to achieve. Because great leadership isn't just supportive, it's predictive. And so far, we've talked about collecting the data, recognizing patterns, but that's just the beginning. Now imagine that our model, our team, or even our career is ready to improve, to grow. In AI, once a model identifies those patterns, it starts iterating towards a goal. But what's the goal? For us, it might be unlocking the full potential of our teams, like reaching, our own version of an Everest Mountain. Right?

For the model, it's about maximizing a target function, an objective that we've defined. And the challenge is that this function is very complex. It has countless of variables, and it's landscape. It's not just this one mountain. It's full of peaks and valleys, and we don't have a clear map, just like we don't have the perfect blueprint for living people. So what do models do in this situation? They search step by step for better terrain. They optimize iteratively, taking one step at a time, asking which direction takes me higher. And it's the same in our careers and leadership journeys. It is called iterative optimization. And when performing iterative optimization, it is very, very important to balance exploitation and exploration. What does it mean to exploit in our context?

In optimization terms, it means that, well, you're standing on a local hilltop in your career, and the next step you'll take is a small one. It's a safe one. It's something that deepens the current expertise that you have or reinforces your known strength. It's a comfortable place to be, but if you only take this kind of steps, you will definitely miss out the Everest of your career. And to reach that higher peak, you just have to take bolder steps, steps of exploration. These are riskier because you might take a wrong turn, but only by balancing exploitation with exploration can we truly unlock the team's potential and our own. And I know some people, some people like to reflect and and set goals around new year.

I don't know if you can relate to that, but for me, it's always it always happens around the International Women's Day. That's when I sit down, reflect on the past year, and set goals for the one ahead. And I have this personal tradition. At least one of my goals has to give me chills just thinking about it because it's so way outside of my comfort zone. It's like I'm picking out the mountains that I want to climb this year, and they become my compass throughout this year, guiding my next steps. But the truth is the truth is that I don't have a map. As we said, we don't have this map. The the this function is so complex. These are just the peaks that I think are out there. But what about the ones that I don't even know exist yet? Right?

Sometimes the most meaningful summits are just the ones that we didn't even plan for, and that's why we need to stay open and flexible enough to say yes when new opportunities just come along. And now let me ask you, what is a key indicator for intelligence? You can even share in the chat if you like. A key indicator for intelligence. Don't be shy. Okay. So, one of the strongest indicators for intelligence is the ability, okay. Someone wrote, accepting that it will go wrong. Humor. I like that. Asking questions. Great. All true. And one of the things in the context that, we share here, one of the key indicators is the ability to learn. Yes. Learning and improving improvement over time and continuous learning. Those are all key indicators. At the heart of every AI model lies some form, obviously, of learning. And, anyone who ever deployed a machine learning model into production knows this. If you train a model, you launch it, and then you just leave it untouched. Right?

You train it, you launch it, untouched, you haven't changed anything. What will happen to its performance over time? Again, you can share in the chat. We train the model, we deployed it, we launched it, and we don't touch it, we don't change anything in the code, what will happen? What will happen to its performance over time? Well, it it will not great. So, yes, it will not it will not remain great for long because without adaptation, without updating it with fresh relevant data, its performance will definitely degrade over time. It will become, as you wrote here, stale, misaligned, and eventually ineffective. And the same is true for us and for our teams because we have to continue and learn, and learning isn't just a nice to have or something that we squeeze in when there is a gap in the spring.

No. It needs to be structured, intentional, like a habit, a practice. And just recently, my husband asked our oldest son if he wanted to do a workout with him. And my son just replied, no, dad. I'd rather train my brain. Now beside melting with pride and saying, yes. That's that's my boy for sure, I realized that he's actually onto something because just like physical muscles, we all have this mental and professional muscles that we need to to train. And the only way to maintain that, let alone improve our capabilities, is to make learning a consistent routine. Okay. Easier said than done. How? How do we do that? Right? And this brings us back to a key skill that we mentioned earlier, and I don't know if you can guess what, skill that is, but I'm referring to pattern recognition.

Can we identify what helps each person on our team to stay motivated and engaged? And on the other hand, what triggers avoidance? And and that mix what's working today might not work tomorrow, so we need to adapt it from time to time as life circumstances change. I can share about myself that there was a time when I could set aside an hour a week and prepare a pile of papers and dive into this stack of technical articles and just learn deeply through focused reading, and that doesn't work for me anymore. These days, I get more value by converting that material into audio, well, 10 to notebook I am, and listening to it like a podcast during my morning commute. That better fits my life now, and, of course, it's highly personal. I had an employee who would block out time on their calendar to improve hands on skills whenever they needed to, to learn new technology.

But despite having that hour scheduled as a routine, avoidance always kept creeping in. When this hour came in the schedule, it started avoiding actually doing it. And what helped? Preparing everything before that hour. The tutorial they wanted to follow, the data they would work in, the development environment all set up. So when that hour starts, it's go time. No excuses. Fully pro productive and onto it. But another person that I mentored, well, found success in in a very different way. She scheduled the learning hour with a colleague. Working together, consulting and sharing ideas, motivating each other, that made all the difference. And these examples just show how important it is to personalize and structure your learning routine. So invest the time and help each person on your team to find their rhythm for learning because ongoing growth, it's well, that's not optional. It's essential.

A team that learns constant constantly just doesn't just keep up, they get ahead. So what does all this mean for us as technical leaders? It means that we can improve our leadership as an algorithm, a process of iterative improvement by gathering data and feedback, but also make it actionable. Identify patterns and take proactive actions ahead. Balance between the familiar zones, but also expand your comfort zones. And foster routines for continuous learning. These aren't just techniques for training AI models. They are the foundation of growth for both ourselves and the teams that we lead. And we've talked about algorithms, but ultimately, we work with people, and the human factor is critical. Leadership isn't just about rigid algorithms. We have another very powerful tool at our disposal, and that's emotion. And when combined with methodical and structured approach, we unlock so much more of our team's potential.

The algorithm, it's it's just a backbone. The heart of leadership is still human. So I invite each of you to reflect. What intentional methodical actions can you start taking today to drive iterative improvement? What's your leadership algorithm? I'd love to hear, so feel free to share and tag me, and I invite you to scan this QR code and connect. And if along the way you realize that you already have some kind of a leadership algorithm of your own, just remember, like models need refreshing to stay effective and relevant, so do we. So I invite you to consider what's your next version update. What process needs a refresh? Disrupt the autopilot of the day to day to level up to an upgraded version of yourself and your leadership. Thank you.