Sometimes we forget the humans behind the tech in our ever busy world. DSF is fortunate enough to know some incredible tech leaders across the world and has the privilege of hearing them present at our events. That being said, our Speaker Spotlight sets the stage to get to know our speakers on a more personal level and connect them with our growing community. Read the mini interview below!
A bit about Anton:
Anton Nazaruk is CTO at Cloud Combinator, where he helps organisations build production AI systems across MLOps, distributed training, cloud infrastructure, and agentic workflows. He started in chemistry and scientific research before moving into data science, quantitative analysis, and AI/ML engineering – a path that shaped his practical, evidence-led approach to building AI systems that work beyond the demo stage.
How did you start out in your tech career?
I started my career in chemistry, where I worked with a lot of experimental data. I was fascinated by how raw measurements could be structured, analysed, and turned into meaningful insight. Over time, I realised I was becoming more passionate about the data side of the work than the chemistry itself, which naturally led me toward data science, machine learning, and cloud-based data systems.
What are the signs of success in your field?
For me, the clearest sign of success is when an AI system stops feeling like a separate tool and becomes a natural part of how people work.
It is one thing to build a good demo. It is another thing to see a team actually change how they operate because the system saves them time, gives them better information, or helps them make decisions faster. That is when you know the work has moved beyond novelty and started creating real value.
On the technical side, success means the system is reliable, measurable, and understood. The team knows what it is doing, how well it is performing, what it costs, when it fails, and how to improve it. There is proper evaluation, monitoring, and governance.
But personally, the most rewarding part is seeing people trust something you helped build – not blindly, but because it consistently helps them do their job better. For me, that is the real sign of success: AI that improves how people operate, without adding unnecessary complexity.
What is the best and worst thing about your job role?
The best thing is working at the intersection of AI, cloud, data, and real business problems. The variety is huge. One day might be about distributed model training and GPU infrastructure, another about MLOps, another about agentic workflows, governance, or helping a team understand how to move from experimentation to production.
I enjoy that mix because it is not just about technology. It is about turning technical capability into something that actually works for people and organisations.
The hardest part is that, as you become more senior, the role inevitably becomes more commercial. In some ways, every senior technical role becomes a sales role: you are constantly selling ideas, priorities, architecture decisions, timelines, trade-offs, and sometimes the value of doing things properly when shortcuts look easier.
That can be frustrating because I still enjoy the deep technical work. But it is also part of the job. Good technology does not create impact by itself – people need to understand it, trust it, fund it, and adopt it. So the challenge is balancing technical depth with the communication and commercial side needed to make the work happen.
What can you advise someone just starting out to be successful?
Learn by doing. There has never been a better time to start building, because modern AI tools have made learning and experimentation much more accessible.
You do not need to wait until you feel ready or until you understand every concept perfectly. Pick a problem, build something small, break it, fix it, and keep improving it. That cycle teaches you far more than passively watching tutorials.
The fundamentals still matter: programming, data, systems, cloud basics, testing, and communication. But the fastest way to learn them is through real projects. Build a simple app, automate something boring, analyse a dataset, deploy a small model, or use an AI tool to help you understand unfamiliar code.
The people who grow fastest are not the ones who know everything at the start. They are the ones who are willing to try, learn in public, ask better questions, and keep shipping.
How do you switch off?
I try to create distance from screens and constant input. That usually means walking, exercise, spending time with people outside of work, and doing things where I am not constantly consuming information.
I am not perfect at switching off, especially because AI is moving so quickly and there is always something new to read or test. But I have learned that being permanently “on” does not make you better. It usually makes your thinking worse.
Good technical and strategic work needs space. Sometimes the best ideas come when you stop forcing them.
What advice would you give your younger self?
I would tell myself to focus earlier on leverage.
Early in your career, it is tempting to measure progress by how busy you are or how many things you can do yourself. But real growth comes from building skills, systems, relationships, and judgment that compound over time.
I would also tell myself not to wait until I felt completely ready. Some of the most valuable growth comes from stepping into rooms where you are slightly uncomfortable – speaking, leading, taking responsibility, building in public, or working on problems just beyond your current level.
You do not need to have all the answers before you start. You need enough discipline to learn quickly, enough humility to correct course, and enough resilience to keep going when things are messy.
What is next for you?
Career-wise, I am focused on helping organisations move from AI experimentation to production AI systems that are reliable, governed, measurable, and genuinely useful.
A lot of companies have now done the first wave of AI pilots. The harder and more interesting challenge is the next stage: building the infrastructure, workflows, evaluation, observability, and governance needed to make AI work safely and consistently at scale.
At Cloud Combinator, that means continuing to build around production AI, MLOps, agentic workflows, distributed training, and cloud-native AI platforms.
I am also increasingly interested in building products that genuinely help people operate. The most exciting products are the ones that remove friction, save time, improve decisions, or make complex life feel simpler.
Personally, I want to keep contributing to the data and AI community through speaking, mentoring, and sharing practical lessons. I think the AI space needs less hype and more honest conversations about what works, what breaks, and what teams should be careful about.
If you could do anything now, career-wise or personally what would you do? Why?
I would build more tools and products that genuinely help people — not only in their work, but also in their everyday lives.
A lot of technology is built around productivity, automation, or business efficiency, and those things matter. But I am increasingly interested in products that reduce friction in a broader sense: helping people understand information, make better decisions, manage complexity, learn faster, stay organised, or get support at the right moment.
That is where I think technology, and especially AI, can become genuinely meaningful. Not just by making companies more efficient, but by making people’s lives a little easier, clearer, or less overwhelming.
So if I could focus on anything, it would be building useful, human-centred tools that combine strong engineering with real empathy for the problems people face.
What are your top 5 predictions in tech for the next 5 years?
1. AI will become part of everyday workflows, not a separate destination
Right now, many AI tools still feel like something you go to separately: open a chatbot, ask a question, copy the answer somewhere else. Over the next five years, AI will become much more embedded into the tools people already use.
The most valuable AI will not always be the most visible. It will sit inside operations, reporting, customer support, finance, software delivery, and decision-making workflows, quietly removing friction and helping people move faster.
2. LLM progress will plateau, and attention will shift toward world models
I think we will start to see diminishing returns from simply making language models bigger. LLMs will keep improving, but the dramatic jumps people became used to may become harder to achieve through scale alone.
The next major step will be models that understand and simulate the world more deeply: systems that can reason across time, space, actions, cause and effect, and physical or business environments. In other words, not just models that predict text, but models that can build useful internal representations of how things work.
That shift matters because many real problems are not just language problems. Robotics, scientific discovery, logistics, operations, planning, and simulation all need models that can reason about systems, not just generate fluent answers. LLMs will still be important, but they may become one part of a broader architecture rather than the whole story.
3. Smaller teams will be able to build much bigger things
AI will make small teams far more capable. A few strong people with good judgment, good tools, and clear focus will be able to build products that previously required much larger teams.
That does not mean everyone becomes a solo founder overnight. But it does mean the advantage shifts toward people who can move quickly, make good decisions, and use AI to compress the distance between idea, prototype, and product.
I think we will see more small teams building serious software, more niche products becoming viable, and more individuals able to create things that used to require a full department.
4. Data quality and context will matter more than model choice
As models become more accessible, the biggest differentiator will be the quality of a company’s data and the context around it.
Organisations that understand their data, where it comes from, who can access it, what it means, how reliable it is, and how it connects to business processes, will get far more value from AI than those simply chasing the newest model.
5. Technical judgment will become more valuable, not less
AI will make it easier to generate code, content, analysis, and prototypes. But that does not remove the need for strong technical people. It actually makes judgment more important.
The bottleneck will shift from “can we produce something?” to “is this the right thing to build, is it correct, is it safe, can it scale, and does it actually help?” People who can ask those questions and make good trade-offs will become even more valuable.
Thank you to all our wonderful speakers for taking part in our Speaker Spotlight!
Want to become a DSF Speaker? Apply here!