ClickHouse in the Wild: Architecture Stories from Production

Join us in Delhi for yet another exciting meetup. This event brings together engineers, architects, and data leaders to explore how modern teams are building scalable, real-time analytics platforms in the cloud. Speakers will share practical architectures, production lessons, and real-world insights from high-volume, performance-intensive workloads powered by ClickHouse. Connect with the local data community and discover how teams are turning streaming data into fast, actionable intelligence.

Don’t miss out! RSVP and secure your spot!

🗓️ Agenda:

• 10:00 AM: Registration & networking
• 11:00 AM: Welcome & opening
• 11:10 AM: Talk 1 - High Speed, Low Latency IoT Logging in ClickHouse by Ashok Vishwakarma, CTO, T9L
• 11:40 PM: Talk 2 - Beyond CDC: Designing a Point-Read ETL Architecture for ClickHouse - How We Migrated 150M+ Rows and Built a Real-Time Analytics Platform by Abhishek Dasgupta, AI Engineer, Zykrr
• 12:10 PM: Break
• 12:20 PM: Talk 3 - TBD
• 12:40 PM: Talk 4 - TBD
• 12:55 PM: Closing Remarks
• 01:00 PM: Lunch & networking

If anyone from the community is interested in sharing a talk at future events, complete this CFP form and we’ll be in touch.

🎤 Session Details: High Speed, Low Latency IoT Logging in ClickHouse
Description: Managing a distributed solar energy grid requires an observability layer capable of handling massive streams of telemetry data from thousands of IoT-enabled sensors. When a solar provider needed to monitor and support their grid infrastructure in real time they faced the classic challenge of high-velocity data ingestion. Traditional databases often buckle under the pressure of sub-second ingestion requirements at this scale particularly when dealing with the dense time-series metrics typical of energy production. This session provides a deep dive into a real-world case study exploring why ClickHouse was selected as the engine to power this IoT logging architecture and how it was implemented to achieve required performance benchmarks.
Speaker: Ashok Vishwakarma, CTO, T9L

🎤 Session Details: Beyond CDC: Designing a Point-Read ETL Architecture for ClickHouse - How We Migrated 150M+ Rows and Built a Real-Time Analytics Platform
Description: At Zykrr, we process millions of survey responses linked to campaigns, participants, schedules, and dynamic survey schemas. While ClickHouse delivered the analytical performance we needed, building a scalable ingestion pipeline proved far more challenging than querying the data. This talk explores three architectures: an initial ClickPipes-based approach that struggled with large-scale joins, a CDC-based design using Debezium that lacked the flexibility required for dynamic schemas and campaign-specific analytics, and the Point-Read ETL architecture we run in production today. In this model, Kafka carries only entity identifiers, consumers fetch and join data in PostgreSQL, transform records into a denormalized analytical format, and write directly to ClickHouse. I’ll also share lessons on schema evolution, failure recovery, real-time processing, and how the same architecture migrated over 150 million records from PostgreSQL to ClickHouse without downtime or data loss.
Speaker: Abhishek Dasgupta, AI Engineer, Zykrr
I'm Abhishek Dasgupta, an AI Engineer at Zykrr working at the intersection of AI, data, and infrastructure. Over the past few years I have worked at 4+ startups, I've built production-scale AI applications, real-time analytics platforms and distributed data pipelines. I'm particularly interested in how AI systems, data platforms, and infrastructure come together to solve real-world problems at scale.