Episode Summary

In this episode, we dig into the hidden layer of AI that few talk about: data infrastructure.
Arturas, who leads data platform innovation at KAYAK, breaks down what it really takes to build scalable, efficient, and secure data foundations that power modern AI systems. From cloud migrations and data lake evolution to retiring legacy systems and cutting S3 costs by half, he shares practical insights on what makes data engineering the real backbone of intelligent systems.
This episode is a deep dive into the balance between innovation and optimization and why even the smartest AI is only as good as the data infrastructure behind it.

Key Takeaways
  • Scalable AI starts with clean, efficient, and well-governed data systems.
  • Cloud migrations and in-house data platforms are reshaping enterprise data strategies.
  • Optimizing infrastructure can drastically reduce operational costs (like S3).
  • Tools like Trino, Snowflake, Spark, and Flink define the modern data stack.
  • Future-ready data architecture focuses on performance, security, and collaboration.
Guest Insight

Arturas leads the development and innovation of KAYAK’s data platform, driving large-scale transformations across cloud architecture, data engineering, and AI infrastructure.

With hands-on expertise in technologies like Trino, Snowflake, Spark, Flink, and Kubernetes, he’s led major initiatives that improved scalability, security, and cost efficiency, including retiring Hadoop and streamlining multi-cloud operations.

A passionate advocate for collaboration and innovation, Arturas focuses on building data systems that not only support AI but also accelerate its real-world impact across industries.

Resources