Trino by Starburst - Bangalore Meetup Group

Agenda
09:30 – 09:40 | Key note from Razorpay CTO / VP Data Engineering
09:45 – 09:55 | Key note from Starburst SVP of Engineering Jitender Aswani
10:00 – 10:30 | Trino advancements over the last 12 months | Ashar Hassan, Senior Software Engineer at Starburst Data, Maintainer of Trino
10:30 – 11:00 | Driving Insights at Scale: Trino-Powered Analytics for Adobe Advertising and Learning Manager Workloads |
Sunil Mandowara, ,Principal Scientist & Architect ,Asish Sahoo AI BigData SAAS Architect, Adobe
11:00 – 11:30 | ☕ Coffee Break
11:30 – 12:00 | Building a Self-Serve Data Platform with Trino at Fintech Scale |
Gaurav Singh ( Lead Data Engineer), Anuj Jain ( Staff Data Engineer), Razorpay
12:00 – 12:30 | Hybrid Serving Architectures: Leveraging HTAP design principles to scale Multi-Tenant SaaS Analytics | Ashish Kumar, Lead Data Architect, 6sense
12:30 – 01:00 | Lakeside AI with Starburst: Turning Governed Data into Enterprise Value | Satya Dixit, Senior Resident Solution Architect, Starburst
01:00 – 02:30 | 🍽️ Lunch & Networking
-------

Trino advancements over the last 12 months, Trino
Over the past year, Trino has continued to evolve as a powerful, distributed SQL engine, delivering significant advancements in performance, scalability, and reliability.
In this session, you'll hear from Trino contributors about key innovations from the last 12 months, including major architectural and performance optimizations, enhanced stability across large-scale deployments, and deeper integrations with modern data platforms and lakehouse ecosystems.
The talk will also dive into a few of the most impactful improvements in greater technical detail—sharing real-world benefits and lessons from the community—before concluding with a forward-looking view of what’s coming next for Trino

Driving Insights at Scale: Trino-Powered Analytics for Adobe Advertising and Learning Manager Workloads, Adobe:
We run Trino as the analytics engine for large-scale, multi-tenant advertising and learning workloads, where queries routinely scan hundreds of tables and terabytes of data, while serving high-concurrency dashboards and scheduled enterprise reports from the same cluster.
In this talk, we share concrete, production-level lessons from operating Trino at scale: cost-based optimization, dynamic filtering, predicate and column pushdown, fault-tolerant execution, and resource-group-driven workload isolation. We walk through real query patterns from ad-tech and learning analytics wide fact tables, skewed joins, time-windowed aggregations, and bursty traffic, and how we tuned Trino to keep latency predictable and costs under control.

Building a Self-Serve Data Platform with Trino at Fintech, Razorpay
At Razorpay, India’s leading fintech platform, we process millions of transactions daily. Behind the scenes, Trino is the backbone of Razorpay’s data ecosystem, powering BI dashboards, merchant and finance reports, invoicing, ad-hoc analytics, ETL pipelines, and production APIs.
In this talk, we share our five-year journey of transforming Trino from a single-cluster query engine into a full-fledged data platform serving 1,000+ daily users and handling 200K+ queries per day across 13 clusters and 3 geographic regions.
What you’ll learn:

• How we implemented workload isolation using Trino Gateway to prevent ETL jobs from impacting interactive queries
• Our custom Kubernetes-based auto-scaling approach that reduced costs by 40% while maintaining performance
• How we built per-query cost tracking and business-unit–level chargebacks that changed user behavior overnight
• The guardrails we enforce at scale, including mandatory partition filters, scan limits, and user quotas

Whether you’re running a single Trino cluster or scaling Trino at the enterprise level, you’ll walk away with practical patterns and battle-tested best practices from operating Trino in production.

Hybrid Serving Architectures: Leveraging HTAP design principles to scale Multi-Tenant SaaS Analytics, 6sense
This session revisits the long-standing debate between Data Lakes and Data Warehouses through the lens of modern enterprise requirements and HTAP (Hybrid Transactional and Analytical Processing) design principles. It explains how traditional architectures evolved to separately handle transactional and analytical workloads, and why this separation introduced data duplication, latency, and operational complexity. By examining schema-on-read vs schema-on-write, governance, performance, and cost trade-offs, the discussion sets the context for why unified architectures are becoming critical in today’s real-time, analytics-driven systems.
In this session you'll learn how 6sense is solving this problem with trino at the heart of their lakehouse engine operating in tandem with other OLTP systems

Lakeside AI with Starburst: Turning Governed Data into Enterprise Value, Starburst
AI is everywhere, but real enterprise value is still bottlenecked by fragmented data, brittle pipelines, and weak governance. This talk introduces Starburst’s data platform for apps and AI and its Lakeside AI approach, built on an Icehouse architecture with Apache Iceberg and Trino to bring AI to your data instead of forcing large-scale migrations.
We’ll walk through the Starburst AI feature suite—Starburst AI Agent, AI Workflows, AI Search, AI SQL Functions, AI Model Access Management, and AI-powered autotagging—and show how they unify natural language access, vector search, and LLM-powered transformations on governed data across hybrid and multi-cloud environments.
The talk will also provide a sneak peek into Starburst's key upcoming AI capabilities.

Want to win our super-cool Community T-shirt?
Sign up for our newsletter for a chance to win!