Santosh Kumar Radha

Santosh Kumar Radha

Santosh Kumar Radha

A theoretical physicist who builds open-source infrastructure for software factories.

Session

Workshop

Date

Monday, November 30

Level

Intermediate

// AI Engineering Skill /

Building agentic systems

// 01 / Bio

Santosh Kumar Radha is Co-founder and CTO of AgentField, where he builds open-source infrastructure for autonomous software, including CodeAF, an open-source coding harness built for open models. His work focuses on making AI systems that build software consistent, reliable and cost-efficient at scale, on models and tools that teams own.

Previously, he led product and engineering at Agnostiq, where he built Covalent, an open-source AI infrastructure platform for workflow orchestration that played a central role in Agnostiq’s acquisition by DataRobot. Santosh holds a PhD in theoretical physics and has published research on reasoning models and agentic AI systems.

// 02 / The session

Software factory architecture: the engineering skill after coding agents

Abstract

Coding agents changed how individuals write code. Most teams still use them one session at a time: a task per chat, a few terminals open, and an engineer moving between them to prompt, wait, review and approve. Output scales with how many agents one person can keep track of.

Software factories get past that limit. Agents work continuously and in parallel, and people step in at a few points placed by design. A factory is an architecture, and every architecture is a set of trade-offs across quality, cost, speed and risk. In this workshop you’ll learn how software factories work from the inside and how to design one around what your team values, so it ships whole features and systems.

What you’ll learn

How a software factory differs from a crowd of agent chats

Where to place human checkpoints, and what the factory does between them

How to trade off quality, cost, speed and risk for your own product

AI DEV · BY DEEPLEARNING.AI

Join us at AI Dev

Meet the engineers shipping AI to production, and leave ready to do it yourself.