Lab Director / AI Infrastructure for Large Scale Models M/F

Adoc Talent Management is recruiting a « Lab Director/AI Infrastructure for Large Scale Models » M/F, for his client, a global leader in telecommunications and digital services. You will join a talented and cutting-edge team within a lab specialized in distributed and parallel computing. This innovative environment encourages collaboration with research teams and focuses on tackling challenging and advanced topics.

You will be a visionary and technically distinguished Lab Director leading the Paris Distributed & Parallel Computing Lab. In this central role, you will drive innovation and breakthroughs in large-scale AI infrastructure, delivering high-quality technologies into real-world applications. As the technical leader of the lab, you will guide both strategy and execution in one of the most advanced AI research directions.

• →Strategic Leadership: Define and lead the lab’s R&D strategy in AI infrastructure for large-scale model training and inference. Identify core research problems and long-term innovation goals.
• →Technology Innovation & Delivery: Lead the team in cutting-edge research and engineering in areas such as distributed parallel computing, automatic parallelization, graph scheduling, and numerical computing. Ensure high-quality technology delivery and practical integration.
• →Team Leadership: Build, manage, and mentor a top-tier team of researchers and engineers. Foster a high-performance, innovation-driven culture.
• →Cross-Team Collaboration: Collaborate closely with the headquarters and global research teams to align efforts and achieve integrated technology deployment.
• →External Influence: Represent the company in global academic and industrial communities. Participate in or lead top-tier conferences (e.g., NeurIPS, ICLR, MLSys, OSDI, SC) and engage in standardization or open-source initiatives.

The lab focuses on AI infrastructure technologies supporting large model training and inference, including but not limited to:

• Distributed Parallelism: Automatic parallelization, pipeline parallelism, tensor parallelism, and communication optimization
• Execution Graph Optimization: Dynamic/static computation graph scheduling and compiler-level graph optimization
• Numerical Computation: Sparse training, mixed precision, low-rank decomposition, dimensionality reduction
• Heterogeneous Computing Systems: System co-design across GPU/NPU hardware, memory, and communication subsystems

The position is based in Boulogne Billancourt (Ile de France, 92) and requires regular on-site presence.

The ideal candidate will have strong technical expertise in one or more of the following areas: large-scale model systems, AI infrastructure, distributed systems, compiler design, or numerical computing.

You should have a proven track record of leading high-impact research or engineering teams of 10+ members, as well as experience designing or delivering scalable AI platforms or model systems in industry or academia.

Prior experience at a world-class tech company, startup, or research institute is essential.

Fluency in English (working proficiency) is required; French language skills are a plus.

Publications in top-tier conferences or journals (e.g., NeurIPS, ICML, ICLR, OSDI, SOSP, MLSys, TPDS), leadership experience in open-source or industrial large language model (LLM) systems, and familiarity with mainstream AI hardware platforms or architectures are highly desirable.