Sat, Nov 21 · 11:00 AM IST
Hello everyone! Join us for an IN PERSON Apache Kafka® meetup on Nov 21st from 11:00AM, hosted by Giniminds in Bangalore!
IMPORTANT!
Please fill out the below form for expense purposes.
https://forms.gle/5RAz7dsVxbHaoemS7
📍 Venue:
Giniminds Solutions Private Limited
738, Sumo Sapphire,
3rd Floor, 15th Cross Road,
J P Nagar 6th Phase,
Bangalore - 560078
***
Agenda:
11:00 - 11:10: Welcome
11:10 - 11:50: Giri Venkatesan, Principal Developer Advocate, Solace
11:50 - 12:30: Sanchitha B, Solutions Engineer - Data Streaming Platform, Giniminds
12:30 - 12:40: Break
12:40 - 13:20: Kumar Keshav, Manager II, Engineering, Confluent
13:20 - 14:30: Lunch
***
💡 Speaker:
Giri Venkatesan, Principal Developer Advocate, Solace
Talk:
How to build a Kafka-compatible engine with no broker disks
Abstract:
Most of Kafka’s operational weight comes from broker-local disks. Replication, rebalancing, capacity planning and cross-AZ traffic all follow from storing data on the brokers. Upstream Kafka is working on this through KIP-1150, and the implementation work is still in progress.
This talk looks at a different route. Nisshi is an open-source (Apache 2.0) engine written in Rust that implements the Kafka wire protocol independently. It is not a fork of Apache Kafka. The broker is a stateless process, and durability is handled by a pluggable storage engine such as PostgreSQL, SQLite or S3-compatible storage.
We will cover how the protocol layer separates from storage, what that separation costs in latency, and where it fits and where it doesn’t. A live demo connects standard Kafka clients without code changes. Then we stop the broker, restart it, and check what survived. We will also say which clients have been tested and what the project does not yet do, including that authentication is off by default.
You will leave with a clear picture of the diskless design space and a way to judge whether it suits your workloads.
-----
💡 Speaker:
Kumar Keshav, Manager II, Engineering, Confluent
Talk:
Powering Real Time Context with Streamhouse
Abstract:
Most enterprise AI agents fail in production because they rely on stale batch data or static snapshots—effectively suffering from "amnesia" about live business operations. In this talk, we explore how to bridge the gap between static LLMs and real-time operational context using Streamhouse. Learn how event-driven architecture continuously streams and contextualizes dynamic enterprise data at query time, transforming reactive LLMs into proactive, state-aware AI agents capable of making accurate, real-time decisions without operating on yesterday's data.
-----
💡 Speaker:
Sanchitha B, Solutions Engineer - Data Streaming Platform, Giniminds
Talk:
From Event Streams to Agentic Systems: Building Real-Time Context and Trusted AI Agents with Apache Kafka
Abstract:
As enterprises increasingly adopt AI agents, the ability to access accurate, real-time and trusted information is becoming critical.
This session explores how Apache Kafka, stream processing and RAG can bring together real-time data and enterprise knowledge to build context-aware AI applications.
Through practical architecture and a live demonstration, we will explore how organizations can move beyond traditional data integration to build scalable, responsive and trustworthy AI solutions that enable faster decision-making, improve operational efficiency and unlock greater business value.