Mistral AI
Research Engineer, Code Agents Infra
About this role
Mistral seeks a Research Engineer to build and operate large-scale infrastructure for agentic AI models, including sandboxing platforms for executing millions of concurrent untrusted code environments, synthetic data generation pipelines, and distributed training systems. You'll optimize systems across hybrid multi-cloud clusters while ensuring security, low-latency execution, and high availability for agent training and evaluation workloads.
What you'll do
- Design and operate a high-throughput sandboxing platform executing untrusted LLM-generated code across 1M+ isolated concurrent environments
- Architect and scale synthetic data generation pipelines for agent trajectories, rollouts, and self-play data collection
- Optimize agent training codebases and distributed runtimes (PyTorch, Ray, SLURM/Kubernetes) to minimize overhead and improve GPU utilization
- Reduce sandbox cold-start times to sub-second levels using warm pools, snapshot/restore, and optimized image delivery
- Implement Kubernetes custom controllers and queuing systems to route evaluation and execution tasks across diverse hardware fleets
- Ensure strict multi-tenant isolation and security for untrusted agent code using container runtimes and network policies
What they're looking for
- Kubernetes and container orchestration (operators, CRDs, custom controllers)
- Sandboxing and virtualization technologies (gVisor, Firecracker, WASM, CRIU)
- High-performance systems programming (Python, Go, C++, or Rust)
- Distributed systems and data pipeline architecture (Ray, Spark, custom queues)
- Linux kernel internals (cgroups, namespaces, container runtimes)
- Task queueing and resource scheduling systems
- ML infrastructure and optimization (PyTorch, SLURM, RL systems)
- Network isolation and multi-tenant security design
Benefits
- Healthcare coverage
- Parental leave
- Retirement plans
- Relocation support
- Wellness programs
- Meal and transportation allowances
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Mistral AI
Mistral AI builds large-scale machine learning systems and open-weight AI models, supported by infrastructure powering petabyte-scale HPC clusters and enterprise AI platforms. The company is hiring Systems Engineers, Research Engineers, Site Reliability Engineers, and Applied AI Engineers to scale training infrastructure, optimize data systems, ensure platform reliability, and drive customer adoption across industries.
- Website
- mistral.ai
Likely interview questions
- Describe a time you optimized a large-scale distributed system—what was the bottleneck and how did you resolve it?
- How have you reduced cold-start latency in containerized or virtualized environments, and what techniques did you use?