VAST Data announced two new services for the VAST AI operating system aimed at managing AI agent activity and improving model performance through continuous tuning.
The products PolicyEngine and TuningEngine are located within the VAST DataEngine. VAST Data positions these as building blocks for organizations seeking large-scale AI deployments with tighter oversight and more predictable behavior.
AI governance is a growing concern for business leaders across the Asia-Pacific region. VAST Data cites research showing that 70% of organizations see AI risks and opportunities as the most pressing board topic for 2026. They argue that the focus has shifted from building models to ensuring that AI workloads follow rules, decisions are traceable, and performance improves over time.
policy control
PolicyEngine targets the risks that arise when AI agents and workflows interact with corporate data and other systems. Agents can generate new data such as responses, agent-to-agent messages, and event logs, increasing the number of places sensitive information can appear.
Without detailed control over what agents can access and how they communicate with tools, remote data products, and other agents, the risk of data leakage increases. VAST Data also says that systems need more robust workflow logging that can be audited and accounted for.
PolicyEngine is designed as an inline policy enforcement layer that checks accesses, actions, and communications against explicit permissions and context before actions are performed.
It also maintains traces and logs that VAST Data describes as tamper-proof, supporting what it calls a “zero trust operating regime” for AI environments. This is in line with the number of security teams currently handling internal services with validation and logging applied as standard controls.
Continuous tuning
TuningEngine addresses a variety of operational challenges. Many organizations deploy models and improve them through separate cycles of data collection, training, and validation. VAST Data proposes a tighter loop that takes results from the AI agent’s workflow and feeds them directly into model improvement.
In this approach, VAST Data’s AgentEngine provides the agent runtime for the AI operating system. It is described as a serverless environment that coordinates multi-agent workflows, model invocation, and tool usage. The platform has been used to deploy static models, but has now been extended to support learning loops based on telemetry and feedback.
TuningEngine collects results from your agent pipeline and applies curated feedback to improve model performance over time. VAST Data says it supports techniques such as LoRA fine-tuning, supervised fine-tuning, and reinforcement learning.
In the workflow described, a tuning pipeline ingests data, processes it, and suggests candidate models. These candidates can be evaluated and benchmarked within the VAST AI operating system prior to deployment. Deployment can be manual or automated, after which new interactions lead to the next improvement cycle.
platform strategy
The broader message of VAST Data is that operational AI requires stronger guardrails and more integrated feedback models. This is framed as a move to systems that “observe, reason, act, evaluate, and improve” while maintaining security and explainability across workflows within a single environment.
In this strategy, PolicyEngine and TuningEngine are intended to work together. One defines and enforces acceptable behavior and access, and the other uses performance signals and real-world feedback to iterate on model quality.
Jeff Denworth, co-founder of VAST Data, said the announcement reflects a broader shift in the way organizations handle AI applications in production.
“Just as people are constantly learning, tomorrow’s applications must also be learned,” Denworth said. “With the introduction of PolicyEngine and TuningEngine, the VAST AI operating system is now a thinking machine that can be deployed anywhere our customers compute. It is a machine that secures every interaction, learns from every outcome, and delivers the power of AI to every organization.”
VAST Data announced the service at the Forward conference in the United States. PolicyEngine and TuningEngine are expected to be released by the end of 2026.
