Almost any application can benefit from having data available locally – enabling blazing-fast access and optimized write patterns. This talk will walk you through one approach to designing a full-featured sync engine, applicable across a wide range of domains, including front-end, back-end, and machine learning training.
I have been working at Turso for over a year and have been fortunate to be involved in initiatives such as implementing native vector search in LibSQL, rewriting our server to support massive multi-tenancy, adopting deterministic simulation testing, and migrating our infrastructure to AWS Cloud. One of my recent projects at Turso focused on improving sync support - a topic so fascinating that I felt compelled to share ideas and use cases with a broader audience. Before joining Turso, I worked at ShareChat on the ML-ops side, building performant and cost-effective infrastructure for modern recommendation systems. During my university years, I participated in competitive programming contests a lot and greatly enjoyed challenging algorithm and data structure problems.
