Name: Keynote: Building for AI at Scale: Infrastructure, Integrity, and Innovation | T.R. Vishwanath | TestMu 2025
Uploaded: 2025-08-30
Duration: 33 min 38 s
Description: In this insightful keynote session, 𝐓.𝐑. 𝐕𝐢𝐬𝐡𝐰𝐚𝐧𝐚𝐭𝐡, Co-Founder and CTO, Glean, shares his deep expertise on building AI systems at scale, emphasizing the challenges and considerations for enterprises adopting AI.

Test your AI Agents with the all-new Agent to Agent Testing Platform.Learn More

TestMu 2025 Home / Video /

Keynote: Building for AI at Scale: Infrastructure, Integrity, and Innovation | T.R. Vishwanath | TestMu 2025

Testμ

Playlist

About the talk

In this insightful keynote session, 𝐓.𝐑. 𝐕𝐢𝐬𝐡𝐰𝐚𝐧𝐚𝐭𝐡, Co-Founder and CTO, Glean, shares his deep expertise on building AI systems at scale, emphasizing the challenges and considerations for enterprises adopting AI. Drawing on his experience building Glean, a knowledge work assistant, he dives into the importance of integrating AI within heterogeneous data systems and ensuring robust data governance, privacy, and access control.

He explores the difference between building for large-scale consumer platforms like Facebook and AI applications for enterprise environments, stressing the need to design systems that handle vast amounts of data while ensuring reliability and minimal manual intervention. Vishwanath highlights the importance of automated tuning and self-healing systems in AI infrastructure to reduce manual overhead.

The session also discusses practical strategies for dealing with data integrity, including understanding the context and provenance of AI-generated answers, and the essential role of AI agents in automating workflows and improving business operations. Vishwanath emphasizes the significance of testing these agents thoroughly through unit tests, integration tests, and end-to-end testing to ensure they deliver consistent results.

𝐒𝐞𝐬𝐬𝐢𝐨𝐧 𝐊𝐞𝐲 𝐓𝐚𝐤𝐞𝐚𝐰𝐚𝐲𝐬:

𝐀𝐈 𝐚𝐭 𝐬𝐜𝐚𝐥𝐞: Challenges of building enterprise-level AI systems with heterogeneous data and diverse customer needs.

𝐃𝐚𝐭𝐚 𝐢𝐧𝐭𝐞𝐠𝐫𝐢𝐭𝐲: The importance of grounding AI-generated answers in reliable data and understanding the context behind each output.

𝐈𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐚𝐧𝐝 𝐜𝐨𝐬𝐭 𝐦𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭: Strategies for optimizing AI infrastructure costs, particularly when using LLM APIs and handling expensive GPU workloads.

𝐀𝐈 𝐚𝐠𝐞𝐧𝐭𝐬: Leveraging agents to automate workflows and the need for rigorous testing and observability to maintain AI system integrity.

𝐀𝐝𝐨𝐩𝐭𝐢𝐨𝐧 𝐦𝐢𝐧𝐝𝐬𝐞𝐭: Engineering leaders must adapt to the fast-evolving AI landscape by embracing rapid prototyping, iterative development, and continuous learning.

Testμ

Testμ (TestMu) is more than just a virtual conference. It is an immersive experience designed by the community, for the community! A 3-day gathering that unites testers, developers, community leaders, industry experts, and ecosystem partners in the testing and QA field, all converging under a single virtual roof. 😀

More Videos from TestMu 2025

Opening Note by Joe Colantonio