Interview with Mathis Joffre, Co-founder and Lead Engineer at Blaxel
Blaxel is a groundbreaking cloud infrastructure platform, specifically designed to cater to the unique needs of autonomous AI agents. Unlike traditional SaaS applications, Blaxel is purpose-built to address the real-time, distributed, and evolving demands of AI software.
Agent-Centric Infrastructure
Blaxel provides infrastructure optimized for autonomous agents, enabling near-instantaneous responses required for voice and interactive interfaces. Unlike traditional cloud services designed for human users, Blaxel connects AI agents to multiple distributed services (Language Models, APIs, memory systems, and knowledge bases) with minimal latency.
Secure Sandboxes
Blaxel offers secure sandboxed environments that can spin up in under 3 seconds and resume from hibernation in milliseconds, ensuring fast cold starts for agents. This allows agents to operate securely and efficiently, isolated from one another.
Serverless Architecture at the Edge
The platform emphasizes a serverless model optimized for AI workloads, enabling high-density deployment, transparency, and simplified user experience for developers.
Memory, Planning, and Execution Support
Blaxel handles core agent functions like memory management, planning, and task execution to support persistent multi-agent systems capable of complex autonomous workflows.
Scalability and Efficiency
By addressing infrastructure misalignments (e.g., network latency, resource allocation), Blaxel reduces operational burdens and costs for companies building AI agents, facilitating better product development rather than infrastructure wrestling.
The Future of Blaxel
As AI agents become ubiquitous across industries, Blaxel is expected to evolve from being "infrastructure for AI agents" to being "the operating layer" they rely on, handling lifecycle, coordination, and even marketplace interactions.
Mathis Joffre, Co-founder and Engineering Lead at Blaxel, was inspired to build this platform due to the increasing complexity of AI and cloud architectures for next-generation AI use cases. The release of Operator by OpenAI further influenced Mathis's decision to design APIs for agents, as it demonstrated the next logical step in AI evolution: agents deploying code autonomously.
Blaxel's architecture includes ultra-fast microVMs, batch job execution, and a global gateway for routing and fallback. The most common mistake when deploying autonomous agents in production versus dev/test is treating agents like functions-invoked, executed, then forgotten. Instead, agents need to persist context, manage tools, and sometimes react to external signals in real-time, which is often overlooked.
The most transformative feature for agent-first systems, according to Blaxel, is snapshot forking, which unlocks debugging, experimentation, and parallel reasoning patterns that aren't possible in conventional cloud environments. The architecture and philosophy of Blaxel are informed by Mathis Joffre's past engineering experiences, focusing on solving real-world needs and delivering a purpose-built platform.
In the future, agents may manage and reconfigure infrastructure autonomously, which has cultural and security implications, including the need for new abstractions for verifiable autonomy and constrained self-improvement. Blaxel supports APIs consumable by agents rather than humans, motivated by the future of agents consuming infrastructure directly. The use of these APIs is expected to evolve, with agents taking on more responsibilities in infrastructure consumption.
[1] Blaxel Blog Post: Introducing Blaxel [2] Blaxel Blog Post: Blaxel's Agent-First Design [3] Blaxel Blog Post: The Future of Infrastructure is Agent-Native [4] Blaxel Blog Post: Building Scalable Multi-Agent Systems on Blaxel [5] Blaxel Blog Post: Blaxel: The AWS for AI Agents
- Blaxel, designed for autonomous AI agents, enhances performance in voice and interactive interfaces by providing infrastructure optimized for near-instantaneous responses, unlike traditional cloud services meant for human users.
- As AI agents become more prevalent, Blaxel, with its purpose-built serverless architecture, is expected to expand its role beyond "infrastructure for AI agents" to become "the operating layer" they rely on, handling complex autonomous workflows and even marketplace interactions.
