Want to learn how to scale your product efficiently, boost team performance, integrate AI into complex systems, and stay strong under heavy traffic? You’re invited to Highload fwdays'26 — a conference for those who build systems that can handle anything!
Want to learn how to scale your product efficiently, boost team performance, integrate AI into complex systems, and stay strong under heavy traffic? You’re invited to Highload fwdays'26 — a conference for those who build systems that can handle anything!
We’ll talk about scaling, performance, databases in production, and more.
Get real-world cases from Uklon, Silpo, MacPaw, Universe Group, and other tech giants.
On June 6 in Kyiv (offline + online), get ready for up to 10 top-tier speakers, and professional networking with engineers, architects, and DevOps pros.
Moving analytics from the backend to the data pipeline is a critical step for systems where millions of events place an excessive load on production databases. In his talk, «How we stopped building analytics directly in the code: CDC, BigQuery, and the new role of the Data Engineer» Yozhef Hisem will explain how the transition to CDC via Debezium and the use of BigQuery allowed us to remove the analytical load from the backend and simplify the system. You’ll also learn about the trade-offs of this approach—from lag and duplicates to cost and working with data schemas.
When data is scattered across dozens of services, traditional approaches don’t work. In the talk «From Microservice Chaos to a Fast Storefront: How We Built a Catalog on Elasticsearch» Oleksii will explain how they built an aggregated read model for 30 million+ documents, handled 1,500 write RPS, and why Elasticsearch became a key component of the catalog.
In the talk «High-load ≠ high-cost: how to optimize infrastructure without losing reliability», we will explore why the cost of high-load systems often grows due to overprovisioning, excessive use of managed services, unnecessary data movement, and non-transparent SLAs rather than the load itself. The speaker will share approaches to cost optimization as an architectural discipline — from workload profile analysis and bottleneck identification to building unit economics, optimizing traffic, caching, CDN usage, observability, and making informed trade-offs between performance, reliability, and cost.
Grab -10% off with a promo code and secure your spot: DEV.UA_HIGHLOAD
Details & tickets here.
Highload fwdays 2026 — where real talk about load happens.
