Build with AI Kolkata: Be Future Ready

Name: Build with AI Kolkata: Be Future Ready
Start: 2026-04-25T09:30:00+05:30
End: 2026-04-25T16:30:00+05:30
Location: Sister Nivedita University, A Block Seminar Hall

Sister Nivedita University, A Block Seminar Hall, DG Block(Newtown), Action Area I, 1/2, Newtown, New Town, 700156

GDG Kolkata

Build with AI are hands-on events run by the community, for the community, where it focused to use the latest Google AI models and tools in practice and build applications using the latest Google AI products including AI Studio, the Gemini and Gemma family of models, Antigravity and Vertex AI

Apr 25, 4:00 – 11:00 AM (UTC)

0 RSVP'd

Key Themes

AIAI - Vertex AIBuild with AIWorkshop / hands-on session

About this event

Build with AI Kolkata: Be Future Ready is a hands-on, community-driven experience designed to help developers turn ideas into real-world AI applications using Google’s latest technologies. Organized by GDG Kolkata, this event empowers participants to explore and build with cutting-edge tools like AI Studio, the Gemini and Gemma model family, Antigravity, and Vertex AI.Focused on learning by doing, the event brings together developers, builders, and AI enthusiasts to collaborate, experiment, and innovate in an interactive environment. Whether you're just starting with AI or looking to deepen your expertise, Build with AI Kolkata is your gateway to building impactful solutions with Google AI.

When

Saturday, April 25, 2026
4:00 AM – 11:00 AM (UTC)

Where

Sister Nivedita University, A Block Seminar Hall
DG Block(Newtown), Action Area I, 1/2, Newtown
New Town 700156

Agenda

3:30 AM	Check-In	Check In and Registration
4:30 AM	Keynote
4:45 AM	Human + AI: Designing Agents for a Collaborative Future	The future of AI isn’t autonomous, it’s collaborative. This talk explores how to build agents that work alongside humans through thoughtful memory design, feedback loops, and human-in-the-loop systems. Learn how to create AI applications that are not just powerful, but trustworthy, controllable, and aligned with real-world needs.Discover practical architectures for building and scaling agent-based systems, that cuts beyond the hype with real business usecase examples
5:15 AM	Building AI Agents with self-managing memory	The challenge with building AI agents is that they do not have a long term persistent memory. We have to manage their memory explicitly. The better way to scale an AI Agentic workflow is to build an AI agent that can self-manage its memory. The concept is very similar to the virtual memory of computer systems where we have 2 tier memory structure.. In this session we will see how we can create an AI agent that can write/ edit its own memory using Google ADK
6:00 AM	From Raw Data to Executive Storytelling: Building an AI-Driven Analytics Pipeline	This session explores a complete "medium-level" workflow: using SQL for data extraction, Gemini for intelligent data transformation and insight generation, and Power BI for final visualization. I will demonstrate how Gemini can act as the "reasoning engine" in the middle of this stack—finding hidden patterns in SQL results and drafting the executive summary that accompanies the Power BI report.
6:30 AM	From Zero to Revenue: Launching Subscriptions with Firebase Extensions	⁠The session will cover what firebase is & how we can embed extensions in firebase. It will demo one example of e-commerce with firebase extensions including payments
7:15 AM	Plug and Play with Firebase Extensions	Building AI applications and scaling your backend doesn't have to be a manual grind. By leveraging Firebase Extensions, you can outsource heavy lifting—from processing payments to deploying AI chatbots—using pre-built solutions that launch in just a few clicks. Automate your workflow, drastically reduce your codebase, and accelerate your path to production by letting Firebase do half the work for you!
8:00 AM	Lunch & Networking
9:00 AM	The Fastest Site I Ever Built Had No Human-Written Bottlenecks	We've spent years hand-tuning bundle sizes, writing custom lazy-load logic, and chasing green Lighthouse scores — and the web is still slow. The bottleneck was never the tools. It was how long it takes a human to find the right problem. In this talk, I'll take a real React + TypeScript app — deliberately broken with layout shifts, render-blocking scripts, and a leaky re-render loop — and use AI-powered tooling to diagnose Core Web Vitals issues in real time, generate targeted fixes, and apply them live on stage. The Lighthouse score goes up. The audience can clone the repo and reproduce every step. But the deeper story is the workflow shift. AI doesn't replace your knowledge of performance — it compresses the feedback loop from "something feels slow" to "here is the exact component causing 340ms of INP, and here is the fix." I'll show where this works beautifully, where it confidently hallucinates, and how to stay in the driver's seat. The final section connects to in-browser AI: a lightweight on-device model that infers network conditions client-side and adaptively serves assets — no server round-trip, no privacy trade-off. Think of it as the natural sequel to where in-browser AI started. You'll leave with production-ready techniques, one open-source repo, and a clear mental model for where AI fits in a performance-first workflow.
9:30 AM	Building Realtime Voice Agents with LiveKit and Gemini 3.1 Flash Live Preview	Voice assistants usually follow a slow pipeline, First speech-to-text then LLM reasoning, and finally text-to-speech. this process adds latency and strips away tone and conversational flow. In this session, I’ll demonstrate how to build a realtime voice agent using LiveKit and Gemini 3.1 Flash Live Preview, Google's latest realtime model that processes streaming audio directly instead of converting everything to text first. I’ll walk through the architecture of a voice agent that joins a LiveKit room, streams microphone audio to Gemini’s realtime API, and returns generated speech back to the user with very low latency. By the end, attendees will understand how to build their own realtime voice assistants and how this approach is better than traaditional speech to text and text to speech pipeline
10:00 AM	Spec Driven Development with Google Conductor	Spec Driven Development using Google Conductor to use ai more reliably and efficiently
10:30 AM	End Note

Speakers

Nikhilesh Tayal

Google Developer Expert, AI

Amey Nerkar

Srushti Raybhoge

Addepar

Software Engineer 2

Aharna Haque

DronaHQ

Developer Advocate

Narendra Nath Chatterjee

StockGro

Android Developer

Sulagna Ghosh

CreoWis Technologies

Solutions Architect

Rohan Giri

C5i

Data Analyst

Arindam Majumder

Studio1

Co-Founder

Organizers

Atri Das

MaxMobility Private Limited

GDG Organizer

See bio

Sumantro Mukherjee

Red Hat

Senior Quality Enginner

See bio

Sarmistha Ghosh

WITDreamin Ambassador

See bio

Akash Kumar Majumder

SWE @ LTIMindtree

See bio

Build with AI Kolkata: Be Future Ready

Apr 25, 4:00 – 11:00 AM (UTC)

Key Themes

About this event

When

Where

Agenda

Speakers

Nikhilesh Tayal

Amey Nerkar

Srushti Raybhoge

Aharna Haque

Narendra Nath Chatterjee

Sulagna Ghosh

Rohan Giri

Arindam Majumder

Partner

GDG on campus Sister Nivedita University

Organizers

Atri Das

Sumantro Mukherjee

Sarmistha Ghosh

Akash Kumar Majumder

Apr 25, 4:00 – 11:00 AM (UTC)