[in-person] Polyglot day: Inferring AI

The Polyglot Meetup has been fostering a strong and dynamic community of developers, over the last 10 years!
Excited to welcome you to our rung 2 in-person meetup of 2026! our highlight this time is learning through community engagement with new session recipe!
We will go length and breadth of AI and Java with expert-led talks covering Transformers, Agentic AI, AI-enabled content generation, and the evolving business landscape surrounding Java licensing. Let us together learn what are the advances in AI, and how they are influencing software engineering and redefining the way organizations innovate.

THIS IS A FREE EVENT! BUT PRIOR REGISTRATION IS MANDATORY
Please Register here
Event Date :- 21st Nov 2026
Reporting Time: 10.30 AM
Session Time :- 11.00 AM to 4.00 PM
Location :
IBM EGL D Block, 3rd floor, Multipurpose hall
Embassy Golf Links, Off IndiraNagar-Koramangala Intermediate Ring Road,
Bengaluru, Karnataka 560071
Map : https://maps.app.goo.gl/qQU2MHAUP5ZeNvtz7

Track 1 :- 11.00 AM to 11.40 AM
Tea break & Networking :- 11.40 AM to 11.50 AM
Track 2 :- 11.55 AM to 1:00 PM
Lunch Break :- 1.00 PM to 1.40 PM ( Lunch will be arranged )
Track 3 :- 1.45 PM to 3:30 PM
Tea Break & Networking :- 3.30 PM to 4:00 PM

[keynote] Transformers : The ABCD of Modern AI
Speaker : Vibha Marpalle
Every conversation about modern AI starts in the same place: the Transformer. it is how models hold context, stay coherent and power nearly every serious AI system today. So lets skip the surface tour and plunge in. Self-attention, positional encoding and encoder-decoder design: lets take each one apart, then put them back together to see how they turn raw tokens into meaning.

Agentic AI for Java: Root Cause Analysis, Code Fixes, and Validation
Speaker : Dinakar Guniguntala
Debugging a Java application in production is only half the battle. Identifying the root cause is valuable, but translating that insight into a code change, implementing the fix, and validating the result is what ultimately resolves the problem.
This session explores how Agentic AI can connect these traditionally separate activities through an integrated developer workflow. We will introduce Causa, an open-source AI-powered root cause analysis tool for Java applications, and demonstrate how it can work alongside an AI coding assistant to accelerate production troubleshooting and remediation.
Using quarkus-perf, a Quarkus application designed for performance testing and controlled failure scenarios, we will trigger a runtime problem and investigate it using Causa. The demo will show how runtime and application evidence can guide root cause analysis, how your AI powered IDE (Bob, Claude etc) can help a developer navigate the relevant source code and implement a targeted fix, and how the change can be tested to verify its impact.
Along the way, we will examine the roles of specialized AI agents, observability data, tool integration through MCP, and human oversight in this workflow. The goal is to explore a practical approach to AI-assisted Java engineering that connects production diagnostics with code-level problem solving, while keeping developers in control of the decisions and changes that matter.

When AI Becomes Your Technical Writer: Engineering Trustworthy AI Content Pipelines
Speaker : Manisha Choudhary
Generative AI is transforming how engineering teams create technical documentation, tutorials, presentations, and onboarding materials, but fluent output does not guarantee technical accuracy. This talk explores how AI-assisted content generation evolves from ad-hoc prompting into a governed, testable engineering pipeline. Through practical examples, we will examine failure modes such as hallucinations, terminology inconsistencies, and prompt regressions, and demonstrate controls including retrieval-assisted generation, automated validation, quality evaluation, and human-in-the-loop review.

Modern AI: reflect, challenge and reinvent!
Speaker : Gireesh Punathil
In 2011, IBM Watson beat Jeopardy champions with a pipeline of hand-built components on a dedicated cluster. Today, a single, rented, general-purpose model can do far more. lets trace how the premises of AI shifted in fifteen years, from architecture and cost to adoption and governance. We will then challenge the assumptions shaping today's systems and explore what is ripe for reinvention - because the history shows that anything we build today will look primitive tomorrow! You will leave with one question: which of today's designs are laws, and which are waiting to break? And perhaps, a fresh perspective, a radical idea or the seed of your AI innovation! Come with energy. Be curious and challenge everything. Lets learn together!

Java Licensing & Why It Matters
Speaker : Suman Mitra
Join this session for a quick overview of how Java licensing has evolved over the years, the current Java distribution landscape, and what organizations should consider when choosing a JDK. We'll compare popular Java distributions, their licensing models & support options