The team
Built by engineers who got burned by uncontrolled agents
Runta started in 2023 when our founders were building production AI workflows and kept running into the same problem: agents with credentials, no egress controls, and nothing in the audit trail. We set out to build the runtime layer that should have existed from the start.
Who we are
A small team with deep runtime roots
We come from platform engineering, security research, and AI infrastructure. Everyone on this team has deployed AI agents in production and lived through the governance problems we are solving.
Guanlan Dai
Chief Executive Officer
Focused on AI systems design, product strategy, and developer-facing infrastructure. Came to Runta after building production AI workflows and finding that the runtime layer needed to be a standalone product, not a library bolted onto application code.
Marcus Pellegrino
Chief Technology Officer
Background in systems programming, sandbox design, and process isolation for cloud workloads. Has spent years building the kind of runtime infrastructure that distributed systems need to be predictable under load, and brought those primitives to the AI agent problem.
Yuki Tanaka
Head of Security
Security researcher focused on egress control, zero-trust network enforcement, and audit trail integrity. Built threat models for AI code execution environments and designed audit log frameworks with regulated-environment requirements in mind. Leads Runta's approach to tamper-evident logging and policy design.
Mission
Runtime governance should not be an afterthought
AI agents are acquiring credentials, calling production APIs, and handling sensitive data. The security review process is catching up. We exist to give platform teams the runtime controls they need to pass those reviews with confidence.
Founded in San Francisco
Three engineers who kept hitting the same governance wall while building AI workflows in production decided the runtime layer needed to exist as a product, not a pattern.
First design partners
Worked closely with security teams at financial services and insurance companies to define the four governance primitives that became the core of the Runta runtime: isolation, quotas, egress control, and audit.
Angel round and early deployments
Closed an angel round and moved from design partner work into early production deployments with teams in financial services and healthcare. Began building the four governance primitives into a shipped product, not just a proof of concept.
Expanding to regulated verticals
Growing the team and deepening support for regulated industries, including healthcare platforms handling patient-adjacent data and financial services firms operating under SOC 2 and ISO 27001 frameworks.
Add a governance layer before your next security review
Get sandbox isolation, resource quotas, egress control, and an immutable audit log for every agent run. Built for security reviews.