ML Infrastructure Engineer (AI)

Location: Hybrid in Paris or London; relocation package available for work from Paris.

Company

hirify.global is an AI Safety company building a safety, reliability, and optimization layer for AI systems through policy testing, enforcement, and continuous improvement.

What you will do

• Build scalable reinforcement learning and LLM post-training pipelines, including smoke-tuning runs and approach ablations.
• Design data control systems for rollouts, replay, filtering, evaluation, and policy updates.
• Optimize training and inference across networking, memory, compute scheduling, storage, checkpointing, and I/O.
• Investigate how infrastructure choices affect learning dynamics, evaluation quality, model behavior, and training stability.
• Build experiment infrastructure for runs, artifacts, evaluations, dashboards, failure inspection, reproducibility, and cost visibility.
• Develop agentic environments with coding-agent harnesses, browser and tool integrations, runtime sandboxes, repository-aware workflows, and multi-agent orchestration.

Requirements

• Experience designing, building, or maintaining distributed reinforcement learning or post-training systems at scale.
• Familiarity with PyTorch or JAX and proficiency in Python, including concurrency, asynchronous programming, multiprocessing, and performance optimization.
• Ability to debug distributed GPU workloads across CUDA, containers, drivers, NCCL or equivalent communication layers, networking, storage, scheduling, and checkpointing.
• Experience with profiling tools such as PyTorch Profiler, Nsight, perf, tracing, metrics, logs, or custom instrumentation.
• Experience with inference stacks such as vLLM, SGLang, TensorRT-LLM, Dynamo, or custom serving infrastructure.
• Strong ownership and ability to turn ambiguous infrastructure problems into working systems and improve them through feedback.

Nice to have

• Open-source contributions, benchmarks, papers with code, or technical writing related to RL, distributed ML, LLMs, inference, evaluation, or agent infrastructure.
• Experience in high-bar AI infrastructure, research, or model environments.
• Experience owning custom training frameworks, fine-tuning pipelines, trainers, schedulers, checkpointing, data loaders, or performance tooling.
• Experience with agentic coding systems, GPU clusters, Kubernetes, Slurm, Ray, custom schedulers, or cloud GPU orchestration.
• Knowledge of NCCL, UCX, NVSHMEM, RDMA, InfiniBand, RoCE, EFA, Rust, C++, CUDA, or Go.

Culture & Benefits

• Small, focused team working on AI safety and complex infrastructure problems.
• Competitive compensation package including equity.
• Flexible time off and paid time off aligned with local regulations.
• Medical insurance in France, learning and development support, and necessary hardware, tools, and services.
• Covered subscriptions for AI agents and IDEs, plus team off-sites twice a year.

Hiring process

• 25-minute introductory call with HR.
• Take-home test assignment.
• Technical interview with the Head of Applied Research followed by a 45-minute final conversation with the CEO.