Dr Anna Andersson is Head of Data Science and AI at Klarian, where she leads innovation in applying advanced analytics to complex operational systems. With a PhD in Physics and extensive experience across academia and industry, she specialises in using AI, machine learning, and knowledge representation to transform raw data into actionable insight.

At Klarian, Anna’s work focuses on operational efficiency and reliability in critical infrastructure, including the classification of system events, ontology-driven reasoning, and the deployment of knowledge graphs. These methods enable operators to contextualise events such as un-commanded pump stops, understand system-wide risk, and act quickly to minimise downtime.

She has directed projects that integrate data from diverse sources into coherent, interoperable frameworks, ensuring decision-makers at every level can engage with information that is accurate, timely, and relevant. Beyond technical delivery, Anna is passionate about making data science practical and accessible, providing tools that enhance situational awareness and empower engineers to optimise system performance.

An experienced presenter, Anna has shared her work internationally on topics including ontology reasoning, self-service analytics, and the future of data-driven decision support in operational environments.
 

1. Looking back, what was the turning point in your career that shaped your approach to data science and leadership?

For much of my career, I thought of myself as an individual contributor - focused on solving problems and building great solutions. I’ve always really enjoyed the synergy that happens when you’re in an environment where a team can thrive and have fun. As the projects grew larger and more complex, I suddenly found myself leading a team, and realised that fostering that collaborative, energising environment had become my job - and one of the most rewarding parts of it.

Another turning point came with my transition from academia into industry. It shifted my perspective. I began to see success not just in terms of technical achievement, but in whether the solutions I built were actually used and delivering value. That focus on impact, combined with the joy of building teams that thrive, has really shaped how I approach both data science and leadership today.
 

2. You work across multiple domains and teams. How do you adapt to new challenges and ensure your solutions deliver real business value?

When I’m working across different domains and teams, my first priority is making sure we have the right people in the room, but not too many at the same time. It’s often more productive to create space for focused conversations with individuals or small groups. In those early discovery phases, I make a conscious effort to listen deeply rather than jump to solutions. The temptation to ‘solution’ too early is strong, but the kind of information you need to solve complex problems rarely emerges in a single conversation. 

Even if I’ll eventually take ownership of the process that delivers the solution, the problem itself belongs to the stakeholders, so it’s crucial to approach it with humility and curiosity. 

One technique I’ve found really valuable is having someone in the room whose only job is to listen. When you’re leading the conversation, it’s easy to get caught up in your own mental model of what’s happening. A dedicated listener helps ensure we’re capturing the full picture. That patience and structure early on ultimately leads to solutions that are not only technically sound but truly aligned with real business value.
 

3. What has been the most complex analytical or AI project you’ve worked on, and how did you approach solving it?

It’s funny, once you finish a project, all the uncertainty is gone and the path suddenly feels obvious. That leaves me always in the middle of the most complex I’ve ever worked on most of the time. 

My current project involves using knowledge representation through an ontology to integrate data and analytics across multiple systems. The challenge isn’t just technical, it’s conceptual. You have to design a framework that’s expressive enough to capture the nuances of the domain, while still being practical for engineers and analysts to use. It’s as much about aligning people and perspectives as it is about building technology.

My approach has been to start from the use cases and work backwards, understanding what questions need to be answered and what decisions need to be supported. From there, I’ve been iteratively developing the ontology and the integration patterns, validating them with real data and real stakeholders. It’s a mix of deep technical design and careful communication, ensuring the system is both rigorous and genuinely useful.

It’s challenging, but also incredibly rewarding especially seeing how knowledge representation can bridge gaps between data, analytics, and decision-making.
 

4. In male-dominated technical fields, how have you established credibility and influence, and what advice would you give to young women pursuing data science or AI?

I try not to focus on the fact that I’m a woman in a male-dominated field, paying too much attention to that can become a distraction. Like in statistics, you can’t draw conclusions about bias from a single data point. My approach has always been to do my best work and to surround myself with people I enjoy collaborating with.

That said, when I’m in a position of influence whether leading a conversation, interviewing candidates, or selecting speakers for an event I’m very conscious of how the dynamics in the room play out and how inclusive the process feels, and I do try to find metrics to track bias even if that is not always easy.

Early in my career, I worked with a senior colleague who could only accept new ideas if they seemed to come from him. It was incredibly frustrating, but it taught me an important lesson about influence. I learned to ‘leave ideas to soak’ rather than pushing too hard in the moment. If a discussion reached an impasse, I’d suggest pausing and returning to it later and often, by the time we revisited it, the idea had taken root and could move forward. It reminded me that progress isn’t always about whose idea it is, but about creating the conditions for good ideas to thrive.

For young women entering data science or AI, my advice would be to focus on building genuine expertise, seek out allies who value your voice, and remember that influence often comes from patience, persistence, and creating space for ideas to grow.

5. How do you balance deep technical problem-solving with collaboration across cross-functional teams?

For me, it comes down to very standard time management, but I think about it in reverse. Instead of blocking out time for deep work, I unlock time for meetings. My default state is focused, technical work, and I create deliberate openings for collaboration when they’re needed. It’s essentially the same principle, just a shift in mindset one that helps me stay grounded in the technical depth of my work while still being fully present and engaged with cross-functional teams.

6. What personal habits or values keep you motivated, and where do you see the most exciting opportunities for AI and data science in the coming years?

I like to embrace uncertainty whilst working towards understand how things work and why. That mindset keeps me motivated, because there’s always something new to explore or a better way to do things. I also value reliability and clarity. I want the systems I build to be trusted and genuinely useful, not just clever.

I stay inspired by reading research papers and books, and by talking to peers. Running the local PyData chapter in Exeter gives me a great opportunity to connect with others who are exploring similar challenges from different angles, those conversations often spark new ideas or help me see things in a fresh way.

Right now, what really excites me is how large language models are changing the way we interact with data. The barrier to asking good questions of data is getting lower, people won’t need to know SQL or the inner workings of a data warehouse to explore insights. That means many more people across an organisation will be able to think and act in a data-driven way. That shift also raises a big challenge: the more we open up access, the more important it becomes to build strong guardrails to ensure reliability. I’m fascinated by what that balance will look like how we design systems that are both accessible and trustworthy.