Akash Agarwal

Swiggy, India's leading quick commerce platform, serves ~13 million users across 653 cities, with 196,000 restaurant partners and 17,000 SKUs. To handle this scale, Swiggy developed a secure, scalable AI platform processing millions of predictions per second. The tech stack includes Apache Kafka for real-time streaming, Apache Spark on Databricks for analytics and ML, and Apache Flink for stream processing. The Lakehouse architecture on Delta ensures data reliability, while Unity Catalog enables centralized access control and auditing. These technologies power critical AI applications like demand forecasting, route optimization, personalized recommendations, predictive delivery SLAs, and generative AI use cases.Key Takeaway:This session explores building a data platform at scale, focusing on cost efficiency, simplicity, and speed, empowering Swiggy to seamlessly support millions of users and AI use cases.

Swiggy, India's top food and quick commerce platform, serves 680+ cities with 200k+ restaurants and a 500k+ delivery fleet, completing 3.5B+ orders. Powering personalized recommendations, route optimization and real-time decisions for 17M+ monthly users, Swiggy uses Snowflake as its data infrastructure backbone. Discover how they achieved 20% faster query performance for seamless peak demand handling, boosting customer experience and retention. Learn about their in-house text-to-SQL gen AI solution, Hermes, leveraging Snowflake to democratize data access for nontechnical users.

Swiggy, India’s leading food delivery platform, processes millions of messages every second to power real-time recommendations, predictions, order tracking, and personalized user experiences. In this session, we’ll explore the challenges Swiggy faced while managing open-source Kafka and how we successfully migrated to Confluent’s managed Kafka cluster, streamlining operations and significantly improving performance. We’ll also dive into the critical role Confluent Kafka plays in our microservices architecture, with a special focus on the complexities of Kafka consumer canary testing. We’ll discuss why this process is complex and how we uniquely solved these challenges to ensure reliable, efficient service delivery. Finally, we’ll demonstrate how Confluent Kafka enables Swiggy to handle millions of messages per second, empowering real-time analytics, predictive models like SLA predictions, and personalized user experiences at scale. This session will provide valuable insights into Kafka’s central role in modern microservices architectures and how Confluent Kafka supports high-performance, scalable, and real-time data pipelines for large-scale applications.