Trajectory Uses Runloop to Scale Continual Learning Workloads

SAN FRANCISCO, June 10, 2026 (GLOBE NEWSWIRE) -- Runloop, the best enterprise-grade infrastructure platform for the development, evaluation, and scalable deployment of AI agents, announced today that recently launched Trajectory is running production workloads on Runloop.

Continuously training and fine-tuning large-scale agentic models for multiple customer products in parallel requires execution environments that spin up at training cadence and run securely with sensitive customer data. Trajectory does this at scale on Runloop, running 10,000+ burst concurrent Devboxes for training and fine-tuning workloads. Each Devbox is an isolated microVM that spins up in under 500ms, executes the job, and tears down when complete, with credentials never entering the training environment and supporting customer-by-customer isolation.

Founded by Ronak Malde, Arjun Karanam, and Michael Elabd, Trajectory is partnering with AI-native companies including Clay, Harvey, Decagon, Mercor, Rogo, and more, to continuously post-train models on real product usage.

Runloop is the only production infrastructure platform that unifies secure execution, persistent agent coordination, credential and access security, and benchmarking under a single control plane for AI agents. Devboxes provide isolated execution environments for training and evaluation, and Blueprints make those environments reproducible across customer workloads. Supporting Trajectory's workload at scale has improved Runloop's platform, including HTTP/2 multiplexing in the SDK, an expanded Linux image support matrix, and dedicated package registry mirrors for fast initialization.

"Continual learning is a defining workload for the next era of AI agents,” said Jonathan Wall, Founder and CEO, Runloop. “Trajectory is betting that the next generation of AI products will be defined by their training loops, not just their model weights. Supporting their training cadence pushed Runloop harder than any prior workload, and we shipped real platform improvements to succeed alongside them."

About Trajectory

Trajectory is the platform for continual learning, helping companies continuously post-train large-scale agentic models on real product usage. Founded in 2026 by Ronak Malde, Arjun Karanam, and Michael Elabd, the company has raised $15 million in seed funding and brings experience from Google DeepMind, OpenAI, Apple, Meta Superintelligence Labs, Amazon AGI, Scale AI, Stripe, and Figma. Learn more at trajectory.ai.

About Runloop

Runloop is the best enterprise-grade infrastructure platform for securely developing, evaluating and scaling deployment of AI agents. Used by companies ranging from top model labs to startups, Runloop reduces time to deploy from months to hours, allowing developers to focus on their agents, not infrastructure. Learn more at runloop.ai.

Media contact:
Michelle Faulkner
Big Swing
617-510-6998
[email protected]

You May Also Be Interested In

• Martinrea International Inc. Holds Annual General Meeting
• Fennec Announces Results of Annual Meeting
• Nova Leap Health Corp. Announces Results of Annual and Special Meeting of Shareholders