PhD · AI Infrastructure · ML Systems · Distributed Training
Paris, France
Open to work & relocation
I build AI infrastructure for distributed LLM training. I received my PhD in Computer Science from Sorbonne Université in July 2026, through a CIFRE collaboration with EURECOM and Huawei Paris.
My planners are integrated into MindSpore and MindFormers. Across these projects, evaluations cover dense, mixture-of-experts (MoE) and multimodal models, clusters of up to 16,384 NPUs, and models with over 1T parameters. The work also supports production training, including China Telecom’s DeepSeek-MoE deployment and TeleChat3-MoE.
How can we plan parallelism for large-scale training without per-cluster profiling, while retaining portability across models and hardware?
I model compute, memory and communication costs, then use optimisation to plan parallelism, pipeline schedules and recomputation. The systems below also support rapid re-planning after model changes and operator-level strategy refinement.
Can Hybrid-Parallel Planning Support Alternating Model–Strategy Design? Dependency-Keyed Per-Layer Primitives for Re-Planning.
Reuses per-layer primitives during hybrid-parallel re-planning, recomputing only those affected by a model edit.
Median re-planning speedup of ~2.7× across audited model edits, with provably identical plan selection to a full search; evaluated on dense and MoE models using up to 4,096 NPUs.
PRISM: Profiling-Free Symbolic Memory-Driven Strategy Planner for Large DNN Model Training.
Osaka, Japan · Jan 26–29, 2026
Plans hybrid-parallel training without profiling, using symbolic memory models, estimates that account for recomputation, and topology-aware communication costs.
Mean memory-prediction accuracy of 92.0% in DeepSeek3 scaling tests and 95.8% across model families, on up to 1,024 Ascend-910 NPUs; up to 1.43× the model FLOPs utilisation (MFU) of Megatron-LM.
BMPipe: Bubble-Memory Co-optimization Strategy Planner for Very-Large DNN Training.
27th IEEE CLUSTER · Edinburgh, UK · Sep 2–5, 2025
Uses integer linear programming (ILP) to jointly optimise pipeline execution and memory use, accounting for recomputation, load imbalance and preparation bubbles.
1.36× speedup over Megatron-Even on a 16,384-NPU Ascend-910 cluster; up to 1.70× with 1F1B-Interleave; the ILP is solved in under 200 ms at production scale.
H²O: Holistic Hyper-Parameter Optimization for Large-Scale Deep Neural Network Training.
Poster · 31st Euro-Par · Dresden, Germany · Aug 25–29, 2025
A two-level optimiser combines symbolic cost modelling with an ILP to jointly plan hybrid-parallel configurations, layer-to-stage assignments and recomputation.
36.7% speedup over the D-Rec baseline on 128-device clusters; 35.7% MFU on DeepSeek-141B — without profiling on accelerators.
ManuMatic: Strategy Injection for Robust Automatic Hybrid Parallelism in Distributed DNN Training.
22nd IFIP NPC · Nha Trang, Vietnam · Nov 14–16, 2025
Lets users specify parallelisation strategies for selected operators, while a D-Rec-based planner determines a consistent strategy for the rest of the graph.
Up to 2.24× speedup over D-Rec on Mixtral-8×7B with expert parallelism (8 NPUs), and 2.04× on Llama3-8B at 8K context. On Qwen2.5-72B at 32K context (64 NPUs, combined with BMPipe): 1.45× over D-Rec and 1.30× over an expert-tuned plan.
A shared interface for PRISM, BMPipe and H²O: configure a model and cluster, compare training strategies and inspect pipeline timelines.
The interactive demo is temporarily unavailable. Explore the planner implementations below.
MLLM Pipeline Bubble Modeling for Large-Scale Training.
Poster · SCA / HPCAsia 2026 · Osaka, Japan · Jan 26–29, 2026
SCOPE — Symbolic Computation-Memory Optimization for Pipeline Efficiency in Ultra-Scale DNN Training.
DP2E-AI 2025 Workshop · Paris, France · Jun 2025
Devices and methods for generating execution plans for a machine learning model.
WO 2025/020165 A1 · PCT/CN2023/109533 · Huawei Technologies · filed Jul 27, 2023 · published Jan 30, 2025
A load balancing method to improve large model performance.
Patent application · Huawei Technologies · publication pending
A multi-dimension parallelism configuration method for large language model training.
Patent application · Huawei Technologies · publication pending
Advisors, committee members, mentors, co-authors and friends who have supported my research.