Bangalore Apache Kafka®, Data & AI

Hello everyone! Join us for an IN PERSON Apache Kafka® and AI meetup on July 18th from 10:30AM, hosted by Nokia in Bangalore!

📍 Venue:
Nokia L5 Manyata Business Park
1, Silver Fir, L5, Manyata Tech Park Rd, DadaMastan Layout, Manayata Tech Park, Nagavara, Bengaluru, Karnataka 560045, India

***
Agenda:

• 10:30 - 10:35: Welcome and Introduction
• 10:35 - 11:15: ZooKeeper to KRaft on Kubernetes: From Concept to Implementation Nokia - Kaushik Srinivas @ Nokia
• 11:15 - 11:55: Comparison of Resource Utilization Across ZooKeeper and KRaft Kafka Clusters, with a Focus on Migration Overhead - Aldan Basil Brito @ Nokia
• 11:55 - 12:10: Break
• 12:10 - 12:50: The Conductor Doesn't Scale: Choreographing Multi-Agent Systems in Production - Giri Venkatesan - Principal Developer Advocate @ Solace
• 12:50 - 13:30: Real-Time Fraud Detection using Apache Kafka and Flink - Yeshwanth L M @ Confluent
• 13:30 - 14:30: Lunch & Networking

***
💡 Speaker:
Kaushik Srinivas, Principal Technical Specialist, Nokia (Bangalore) https://www.linkedin.com/in/kaushik-srinivas-216098129/

Talk:
ZooKeeper to KRaft on Kubernetes: From Concept to Implementation

Abstract:
This session explores the practical journey of migrating Apache Kafka from ZooKeeper to KRaft within a Kubernetes environment. Beyond the theoretical concepts, the talk focuses on real-world implementation strategies, rollback challenges, key architectural hurdles and the solutions used to resolve complex migration challenges. Time permitting, a live cluster migration demo will showcase the process in action.

-----
💡 Speaker:
Aldan Basil Brito, Senior Technical Specialist, Nokia (Bangalore) https://www.linkedin.com/in/aldan-brito/

Talk:
Comparison of Resource Utilization Across ZooKeeper and KRaft Kafka Clusters, with a Focus on Migration Overhead

Abstract:
This session provides a data-driven comparison of resource utilization between ZooKeeper-based and KRaft-mode Apache Kafka clusters under identical workloads. Rather than theoretical derivations, the talk presents empirical facts and comparative metrics of both footprints. Special emphasis is placed on the temporary resource overhead introduced by the coexistence of ZooKeeper and KRaft pods during migration, concluding with practical strategies to optimize this peak footprint.

-----
💡 Speaker:
Giri Venkatesan - Principal Developer Advocate @ Solace - https://www.linkedin.com/in/girivenkatesan/

Talk:
The Conductor Doesn't Scale: Choreographing Production Multi-Agent Systems

Abstract:
Production multi-agent systems don't fail because the agents are bad - they fail because a single conductor can't scale. We'll trace the failure modes of orchestrator-led coordination, show why event-driven choreography is the right foundation, and prove it live, with a deeper look at where Solace and Kafka each fit.

-----
💡 Speaker:
Yeshwanth L M, Solutions Engineer @ Confluent - https://www.linkedin.com/in/yeshwanth-l-m/

Talk:
Real-Time Fraud Detection using Apache Kafka and Flink

Abstract:
Waiting for a nightly batch job to flag credit card fraud means the damage is already done. In 2026, users expect security to happen in the blink of an eye. So, how do we build systems that catch bad actors before the transaction even clears?

In this meetup talk, we’ll dive into the world of event-driven architectures using two powerhouse open-source tools: Apache Kafka and Apache Flink. We’ll look at a practical, end-to-end implementation where simulated credit card transactions are streamed into Kafka and analyzed instantly by Flink. We will explore how to handle streaming data (like amounts, timestamps, and locations) and how Flink uses stateful stream processing to spot suspicious spikes and anomalies in real time. Whether you're a seasoned data engineer or just curious about stream processing, you'll walk away with a clear blueprint of how Kafka and Flink team up to tackle real-world problems.

-----