Siva Rama Krishna Varma Bayyavarapu

Title of the Talk:
From Autocomplete to Pull Requests: Engineering Guardrails for Autonomous Coding Agents

Abstract :
AI-assisted software development is moving beyond code completion toward autonomous coding agents that can take bounded engineering tasks, modify repositories, run tests, and prepare pull requests for review. This talk examines that shift from a practical software engineering perspective. It discusses where coding agents can help, especially in repetitive and well-scoped workflows such as configuration updates, logging changes, test scaffolding, and small refactoring tasks. The session also focuses on the guardrails needed for responsible adoption, including scoped permissions, policy-based task filtering, controlled tool access, test enforcement, human review, observability, and audit trails. The goal is not to replace engineers, but to reduce repetitive work while preserving accountability, code quality, and production trust.

Bio:
Siva Bayyavarapu is a Distinguished Fellow at SCRS, a Fellow of IETE, and a Lead Software Engineer at DocuSign with over 15 years of experience in enterprise cloud systems, distributed architectures, and large-scale SaaS platforms. He previously worked at Salesforce, where he contributed to building Salesforce Marketing Cloud capabilities.
He has led the design and implementation of high-performance integration frameworks across the Salesforce, SAP Ariba, Microsoft, and DocuSign ecosystems, supporting mission-critical automation for global enterprises at scale.
His work focuses on enterprise integrations, cloud orchestration, event-driven architectures, and AI-driven automation for enterprise platforms. He is also a published author in IEEE and international conference proceedings and serves as a peer reviewer for multiple scholarly venues, contributing to both industry