AI native observability is now a strategic requirement. Gabriel-James Safar, co-founder and CEO of Tsuga, joins Alan Shimel to explain why traditional observability stacks buckle under AI SRE agents and how a bring your own cloud deployment restores control. Furthermore, GJ unpacks Tsuga's approach to governance, sovereignty and open telemetry.

About Gabriel-James Safar

GJ trained as an engineer, became a data scientist and then co-founded a startup that Datadog acquired. Consequently, he brings deep observability chops to a company built for the AI era, and he now leads Tsuga with the same co-founder for a third time.

Why AI native observability breaks the old stack

Alan digs into the product. Tsuga runs inside the customer's own cloud. As a result, telemetry never leaves the customer environment, residency stays under customer control and compliance moves closer to the workload. Meanwhile, Tsuga stays compatible with OpenTelemetry.

Open source storage engines like Apache Arrow, Data Fusion and Ballista slot in as well. Therefore, teams keep full portability across their data lake and business intelligence tools without vendor lock in.

Inside the AI native observability platform

GJ then walks through the AI SRE model. In addition, Tsuga uses a bring your own agent design so customers plug in whatever agent their AI policy approves. Root cause analysis, aggregate lookups and faulty deployment detection run through that agent against the same telemetry.

Explore more AI coverage and the latest TechStrong TV interviews. GJ also covers the recent Series A led by General Catalyst with Singular and DST joining the round.

How AI native observability supports agents at scale

AI SRE agents can fire hundreds of queries at once. Consequently, older telemetry platforms crash under that load. Tsuga aggregates information so agents consume fewer tokens and explore a smaller search space, which keeps AI native observability fast and reliable at agent scale. Learn more at tsuga.com.