​Building AI Agents That Win at Scale

​Hosted by Magicball x Flexprice

​AI agents are moving fast. From being impressive demos to powering onboarding, documentation and multilingual user experiences across real products. But how do you take an agent from a single prompt to something that users trust, teams adopt and businesses can actually run at scale?

​This session brings together founders, engineers and product leads who’ve built AI-native workflows that automate video creation and documentation, shipped open-source multilingual models and speech APIs and architected monetization infra for production-grade agents.

​We’ll unpack the infra and product decisions behind agents that don’t just work, but win.

​Meet the Panel

​Prajwal Prakash | Co-founder & CTO, Clueso (YC 23)
Prajwal is building AI-native workflows that turn raw screen recordings into polished product videos and documentation — no editors, no writers, just clean automation. He’s shipped features that power onboarding, internal enablement, and customer education at scale.

​Abhishek Upperwal | Founder, Soket AI Labs
Abhishek is building India’s multilingual foundation models and real-time speech infra. From Pragna‑1B to Soket’s voice APIs, he’s deep in the weeds of what it takes to make AI agents talk, understand, and scale across diverse languages.

​Nikhil Mishra | Co-founder & CTO, Flexprice
Nikhil is building the monetization platform for AI & Agentic Companies which work at scale. At Flexprice, he’s helping AI companies monetize agents with real-time metering, billing, and entitlements, without duct-taped workflows.

​
In this session, we’ll cover:

• ​How to move from AI demos to real-world adoption

• ​Building infra that supports multilingual, real-time agents

• ​Automating onboarding and documentation with LLMs

• ​The tradeoffs behind model choices, UX, and cost at scale

• ​What breaks when you scale agents and how does one fix it and more

​👀 Who Should Attend?

• ​AI Engineers and Founders building open-source workflows

• ​Engineering Leaders and Enterprise Engineers exploring what’s real vs hype in Agentic AI

• ​Investors actively deploying in engineering systems

​📍 Location: HSR Layout, Bengaluru

​🎟️ 50 spots only – each registration is reviewed to ensure a high-signal, technical audience.

​This event is powered by –

​Flexprice – An open-source metering & billing platform for AI & Agentic companies. Flexprice simplifies usage-based pricing, credits, and enterprise contracts so engineering teams can focus on building product, not billing.
GitHub: https://github.com/flexprice/flexprice

Magicball is India’s largest community of AI engineers, founders, and operators. We host live product demos, give feedback that actually gets implemented, and connect people solving real AI challenges — helping your product reach the right audience. See what others are building, learn what works (and what doesn’t), and move faster together.
Join us at magicball.dev to plug into the action.