Research Engineer, Code Agents Infra
Mistral AI1 month ago
Palo Alto, CA, USAMid Level
Responsibilities
- Design, deploy, and operate high-throughput sandboxing infrastructure for untrusted LLM-generated code.
- Build scalable synthetic code generation, agent trajectory, rollout, and self-play data pipelines.
- Optimize PyTorch training codebases and distributed execution runtimes for rollout speed and GPU utilization.
- Reduce sandbox startup times using warm pools, snapshot/restore technologies, microVMs, and optimized image delivery.
- Implement Kubernetes controllers, CRDs, queueing systems, and resource allocation across hybrid and multi-cloud clusters.
- Enforce multi-tenant process and network isolation using container and sandboxing runtimes.
- Maintain high availability, telemetry, automated self-healing, and on-call support for transient agent workloads.
Requirements
- At least 4 years of experience in systems engineering, distributed systems, cloud infrastructure, or MLOps supporting LLM or reinforcement-learning workloads.
- Experience building high-throughput data processing and generation pipelines for large-scale datasets using technologies such as Ray, Spark, or custom distributed queues.
- Strong Kubernetes and container expertise, including custom operators/controllers, Linux cgroups/namespaces, and Docker image optimization.
- Advanced proficiency in Python, Go, C++, or Rust, with experience profiling and optimizing high-performance ML or backend systems.
- Hands-on experience with lightweight virtualization, container runtimes, or WebAssembly, such as Docker, gVisor, or Firecracker.
- Deep familiarity with task queues, resource schedulers, and low-latency queuing architectures for high-volume short-lived workloads.
- Ability to collaborate directly with AI researchers and turn experimental agent concepts into production infrastructure.
Benefits
- Benefits may include healthcare coverage, parental leave, retirement plans, relocation support, wellness programs, meal and transportation allowances, and location-specific perks; offerings vary by country.
Categories
About Mistral AI
Mistral AI builds foundation language models and full‑stack AI solutions for enterprises, offering APIs, on‑prem deployment, developer tools, and applications. The privately held company, founded in 2023 and headquartered in Paris, partners with organizations in finance, manufacturing, defense, healthcare, and the public sector to co‑create customized systems. It releases open‑source models alongside commercial offerings and makes its models available through major cloud marketplaces.