Tensil Labs

Tensil Labs builds evaluation and observability tooling for teams running machine-learning models in production. The product measures model drift, output quality, and performance against benchmarks so engineering teams can catch regressions before they reach end users. The buyer is a machine-learning or platform engineering lead responsible for models in live systems.

Founded in 2021 and incubated internally, Tensil is the earliest-stage holding in the AI/ML group, serving roughly a hundred-and-a-half engineering teams with revenue in a modest early band. Its trajectory is measured against adoption among production teams rather than near-term revenue, and early retention among teams that wire it into their pipelines has been encouraging.

As an incubated, early-stage company, Tensil consumes a lean shared-services subset—HR, Accounting, and IT—under the Core management-fee tier, keeping its operating cost low while it establishes product-market fit. It draws directly on sister companies Meridian AI and Sablefox as both reference customers and integration partners.

Forward focus is on broadening evaluation coverage and serving as the observability standard across the portfolio's own model deployments.