Bangalore Streams In-Person meetup - May 2026

Hello Bengaluru👋
Join us for exciting discussions in the streaming world with opportunities to network with peers and leaders in the industry.

📅 When: May 30 2026 9:30am - 02:00pm
📌Where: Razorpay, Bengaluru
🗺Directions: https://maps.app.goo.gl/NzBW7ksNfcxUseef7 (Razorpay, SJR Cyber Laskar, Hosur Rd, Adugodi, Bengaluru, Karnataka 560030)

🕒 Schedule :
10:00 am - 10:20 am: Welcome & registrations
10:30 am - 11:15 am: `Real-Time RAG: Can Retrieval Systems Keep Up with Streaming Data` by Lakshay Kumar, GenAI Lead @ Zato
11:20 pm - 12:00 pm: `Anomaly Detection and Streaming at Razorpay` by Narendra Kumar & Anuj Jain, Razorpay
12:00 pm - 12:15 pm: Networking break
12:15 pm - 1:00 pm: `Embedding Kafka at the Edge: SBCs, Dead Badgers & New Deployment Models` by Avinash Upadhyaya & Pavan Keshavamurthy, Platformatory Labs
1:00 pm - 2:00 pm: Snacks & Networking

🎙️Talks:
Real-Time RAG: Can Retrieval Systems Keep Up with Streaming Data
Speaker: Lakshay Kumar, GenAI Lead @ Zato | Author of “from data to world” | Automating workflows
About the talk: Retrieval-Augmented Generation (RAG) has quickly become the default architecture for enterprise GenAI systems - but most RAG pipelines are fundamentally designed for static or batch-updated data.
What happens when the knowledge base itself is constantly changing?
In this talk, we’ll explore the emerging challenge of Real-Time RAG: building AI systems that can ingest, index, retrieve, and reason over continuously streaming data with low latency and high freshness guarantees.
Using practical architectures and production-inspired patterns, we’ll look at how streaming systems like Kafka enable continuously updating AI knowledge pipelines, and where traditional vector-search architectures begin to break under real-time workloads.
The session will cover:
\ *streaming ingestion for RAG systems
* incremental embeddings and vector updates
\ *balancing freshness vs retrieval quality
* event-driven AI architectures
\ *consistency challenges in distributed retrieval systems
* observability and operational concerns for live AI systems

Embedding Kafka at the Edge: SBCs, Dead Badgers & New Deployment Models
Speakers: Avinash Upadhyaya & Pavan Keshavamurthy, Platformatory
About the talk: Kafka powers event-driven systems everywhere—except many edge environments, where diverse protocols, resource limits, and distributed-system constraints create unique challenges. This talk explores what Kafka at the edge should look like: new architectures, deployment models, codecs, and security considerations tailored for connected cars, factories, telco towers, POS systems, and IoT devices. Learn why current technologies fall short, what a SWaP-C-optimized Kafka could enable, and what you can start doing today to prepare your edge streaming stack.

Streaming Patterns in Production: Lessons from Building at Razorpay
Speakers: Narendra Kumar & Anuj Jain, Razorpay
About the talk: Streaming architectures look clean on whiteboards and messy in production. This talk distills the architectural patterns behind Razorpay's real-time data platform — built and battle-tested while processing 5B+ events a day across payments, banking, payroll, and cross-border. We'll cover how to choose between batch and streaming execution, when declarative beats imperative, how to design for multi-tenancy from day one, and where the boundaries between ingestion, processing, and serving make or break a platform at scale. Each pattern is grounded in real production examples — including the design choices that mattered, and the ones we'd undo.