Alexis Lê-Quôc

Off-the-shelf models are easy to deploy, but they are rarely enough to solve complex, domain-specific challenges in production. The key to sustained AI value is not in the models themselves but in the ability to tune, evaluate, and refine those models against your organization’s real-time signals. We are excited to announce that Adaptive ML is joining Datadog to accelerate this vision by combining our deep observability data with their expertise in building specialized, high-performance AI agents.

Adaptive ML is a frontier AI startup whose Adaptive Engine platform enables enterprises to build, own, and deploy specialized agents and models. Rather than depending on one-size-fits-all APIs, Adaptive Engine lets organizations fine-tune open models using reinforcement learning and synthetic data generation, evaluate them against real business outcomes with bespoke AI judges and A/B testing, and close the loop by feeding production signals back into training. The result is a specialized model that your organization actually owns and that gets better with every deployment.

By combining Adaptive ML with Datadog’s unmatched access to real-world infrastructure, application, and security data, we can continue to build specialized agents and models that are genuinely useful in production environments, trained on the actual signals that matter in cloud observability and security. Together, we can push further on the research challenges at the frontier of autonomous operations, moving from agents that investigate and alert, to systems that learn, adapt, and act.

Stay tuned for future updates as we work to bring Adaptive ML’s capabilities into Datadog.