Skip to main content
TALENT247 // SPECIALIST AI EXECUTIVE SEARCH & TECHNICAL RECRUITMENT
ML · Machine Learning

Senior RL & Sandboxed Code Execution Systems Engineer

Paris, France · Full-time · Hybrid

AI Specialism

Machine Learning

Workplace

Hybrid

Location

Paris, France

Compensation

€120k – €170k

annual package

Role Dossier & Specification

REF: SENIOR-RL-SANDBOXE

THE CLIENT & MANDATE

Our client is a $500M+-funded frontier AI company with its primary engineering and research hub in Paris, building custom foundation models engineered specifically for enterprise software development via Reinforcement Learning from Code Execution Feedback (RLCEF). In their training architecture, the bottleneck to smarter coding models isn't just GPU FLOPs—it's executing millions of heterogeneous, stateful multi-container repository builds, tests, and browser/API checks per minute safely and deterministically.

They have retained Talent247 to hire a Senior RL & Sandboxed Code Execution Systems Engineer in Paris. Joining their RL Infrastructure & Execution Environments team, you will own the planetary-scale microVM/container orchestration engine and asynchronous RL rollout pipeline connecting multi-thousand-GPU training clusters with real-time compiler and test-suite reward environments.

KEY RESPONSIBILITIES

  • Architect, scale, and optimize high-throughput sandboxed execution infrastructure (Firecracker microVMs, gVisor, overlayfs/snapshot-restore) running millions of untrusted multi-language repository builds and test suites concurrently.
  • Build low-overhead asynchronous RL rollout engines (orchestrating vLLM/SGLang inference workers, environment steppers, and PyTorch FSDP/Megatron learners) maximizing GPU utilization during long-horizon agentic coding episodes.
  • Engineer hermetic build-caching, deterministic network/clock virtualization, and anti-tampering sandbox isolation so RL agents cannot game test harnesses or leak side effects.
  • Create automated repository ingestion pipelines that turn raw multi-service monorepos (C++, Rust, Java, TypeScript, Python, Go) into reproducible, sub-second-bootable RL environments.
  • Profile and eliminate tail-latency stragglers across distributed storage, container snapshotting, and RPC trajectory streaming.
  • Partner directly with RL research scientists to expose rich intermediate execution traces, coverage diffs, and profiler signals as dense reward inputs.

TECHNICAL ENVIRONMENT & OFFERING

Their infrastructure stack runs on Rust, Go, C++, Python, Linux kernel virtualization (KVM, eBPF, cgroups v2, namespaces), Kubernetes, vLLM, and dedicated GPU + high-core-count CPU superclusters. Working hybrid in central Paris, our client offers €120,000–€170,000 EUR base salary plus substantial equity, comprehensive family health coverage, and full relocation/visa support.

Core Requirements

ESSENTIAL QUALIFICATIONS

  • 5+ years of distributed systems, cloud runtime, or ML infrastructure engineering experience operating at massive concurrency.
  • Strong systems programming mastery in Rust, Go, or C++, alongside solid Python proficiency for ML/RL pipeline integration.
  • Deep hands-on expertise in Linux isolation and sandboxing primitives (Firecracker, KVM, gVisor, namespaces, cgroups, seccomp, eBPF, copy-on-write filesystems).
  • Experience building high-throughput job schedulers, CI/build execution engines (Bazel/Nix/containers), or distributed RL rollout frameworks (verl, Ray, OpenRLHF, or custom equivalents).
  • Ability to debug subtle kernel, filesystem, and distributed concurrency bugs that only surface at million-container scale.

PREFERRED / BONUS EXPERIENCE

  • Direct experience supporting frontier LLM post-training or SWE-agent reinforcement learning runs.
  • Familiarity with vLLM/SGLang continuous batching, KV-cache prefix sharing across tree rollouts, and RDMA/high-speed networking.

Position Telemetry

Location Paris, France (Hybrid)
Department Machine Learning
Employment type Full-time
Workplace type Hybrid
Salary range
€120k – €170k annual package