// about daxa
Daxa is the data control layer for agentic AI. We decide what agents may see and do with enterprise data, and enforce it in the infrastructure.
// our story
Daxa started in 2022 with a simple observation: AI is only as safe as the data it can reach. While most of the industry worked on filtering what models said, we worked on what they could see. We built permission-aware data controls for AI apps and released Pebblo as open source so builders could have them from day one.
Then AI started to act. Agents now read, change and delete data across systems, at machine speed, with no person in the loop. The risk moved from what AI says to what AI does with your data.
In 2024 we rebuilt Daxa to enforce in the infrastructure itself, from network to endpoint, so agents can't route around it. Today Daxa is built into HPE AI Factory and Private Cloud AI, and we work with NVIDIA on agent governance.
AI That Acts Needs A Control Layer.
CC
“Zero Trust and AI Firewalls can't protect Enterprise Agents without reasoning over the data. Daxa saw that very early on.”
Dr. Claudionor Coelho, Jr. · Former Head of AI, Palo Alto Networks and Zscaler
// milestones
●
2022
Founded on data control for AI
●
2023
Pebblo released as open source
●
2024
Rebuilt for infra-native enforcement
●
2025
Gartner AI TRiSM · HPE partnership · NVIDIA Inception
●
2026
On HPE AI Factory · agent governance with NVIDIA
// what we believe
Decide what reaches the model and what the agent may do with it.
Controls that live outside the agent can't be talked out of.
Risk shows up in patterns across steps.
Expertise in LLM security, AI safety research, and evaluating model risks.
.svg)
Expertise in machine learning, LLM evaluation, and building research into production systems.
Expertise in Python, Data pipelines, Gen-AI application frameworks like Langchain
Experienced QA engineer with expertise in Cloud and API testing
Expertise in data security and high-scale data pipeline architectures
