A Strategic Approach to Data Intelligence Success

Posted on 17/10/2023, Martina Ivanicova

Martina Ivaničová is our Engineering Manager of Data Engineering based in our office in Bratislava, Slovakia. In this article she provides insight into our Data Engineering Team.

Every month, people book hundreds of thousands of flights through Kiwi.com, and we handle tens of millions of search requests daily. I, as Engineering Manager of Data Intelligence, trust that data is fundamental to us; without it, we couldn’t provide top-notch travel itineraries and ensure users have a smooth travel experience. For instance, we conduct hundreds of experiments yearly to optimise the user experience with our product and utilise machine learning (ML) models for search results optimization.

The evolution of the industry around data over the last twenty years is fascinating. Increasingly, the preference is to base decisions on data – indeed, one in three employees at Kiwi.com does look atdata insights daily. Personally, data has always been integral to my professional life. My experience ranges from performing analytics on large production databases back in the early days, to developing warehouse systems for large companies, and creating end-to-end solutions for smart buildings—from IoT (Internet of Things) level data collection to prescriptive analytics.

When I joined Kiwi.com three years ago as Data Platform Lead, we were on the verge of embracing the Data Mesh approach. We were among the early adopters of this paradigm. Essentially, Data Mesh applies microservice architecture to data, emphasising domain data ownership and treating data as products. With my expertise in both software engineering within microservice architecture and experience with data warehouses, I was perfectly positioned to spearhead this significant initiative alongside my team. 

Today, while we haven’t yet reached our final destination, I’m pleased with the results we’ve achieved. The number of users utilising data in a self-service manner has doubled, our data satisfaction score has risen by 20%, and we’ve successfully transitioned to a self-service data platform, reducing the workload on the central Data Platform team by 50%.

Our Purpose in the Data Engineering Tribe

Alongside my colleagues in the Data Engineering tribe, the tribe I lead today, we ensure that individuals and computers can access the right data to make decisions that shape the future of travel. We are teams of analytics engineers who specialise in modelling data into business domain data products, making them available for self-service analytics, ML model training, or performance controlling. The objective of our data platform is clear: we strive to offer a reliable platform and tools that make working with data straightforward for both data developers and practitioners, thus democratising it for everyone.

I believe that self-service is the key concept here. Especially with the advancements in GenAI, we are focusing more on what makes us human, delegating mundane and repetitive tasks to computers. I’m of the view that people shouldn’t have to ask other people data questions.

How Do Members Of The Data Intelligence Team Experience Professional Growth?

We’ve established a robust system for performance management. This system provides clear guidelines on the competencies expected for each role, fostering a culture where open and honest feedback is the norm. Promotions are not based on a manager’s subjective opinion; instead, the criteria for professional advancement are transparent and predefined.

Given the myriad of business tasks, finding time for skill enhancement can be a challenge. To address this, we’ve introduced the Data Engineering tribe “learning weeks”. During these periods, team members step away from their regular duties to focus on what they consider as more valuable for their development. This typically involves training sessions, certifications,  hands-on exploration of new technologies, or diving into industry-standard literature.

Collaboration In Time-Sensitive Situations

Although our actions in the Data Engineering tribe might potentially influence revenue or result in missed opportunities and incorrect business decisions, they do not typically compromise the travel experience. Therefore, I find solace in knowing that we do not face the extreme pressures that some other engineers do. That said, we remain committed to achieving engineering excellence and maintaining a blame-free culture. 

When an outage happens we first of all focus on transparent communication of the situation, identifying the impact and finding the quick, even, if not the ultimate fix to the situation. We aim to learn from every incident to prevent its recurrence. To facilitate this learning process, we conduct Post Mortems to analyse each incident and develop strategies to avert similar occurrences in the future.

We are on the lookout for new people to join our Engineering teams. If working at Kiwi.com sounds interesting, look at our open roles and apply.