Machine Learning Engineer - Foundational
Harmattan AI6 months ago
Paris, FranceSenior
Responsibilities
- Design multimodal self-supervised learning architectures and loss functions for paired and unpaired electro-optical and infrared data.
- Manage and optimize distributed training pipelines across multi-node GPU clusters, including mixed-precision training and data loading.
- Develop representation-quality metrics and linear-probing benchmarks for latent-space evaluation.
- Audit EO/IR data lakes and implement cross-attention mechanisms for sensor-feature fusion.
- Collaborate with Data Engineers on ingestion pipelines and with the Edge AI team on high-performance model handoffs.
- Architect fault-tolerant data-pipeline state machines and mediate technical trade-offs between hardware and algorithm teams.
Requirements
- PhD or highly research-focused MS in Computer Science, Machine Learning, Computer Vision, Applied Mathematics, or a related field.
- Minimum 5–6 years of experience for senior levels.
- Experience training and scaling deep-learning vision models, including ViTs and CNNs, from scratch in multi-GPU and multi-node environments.
- Experience applying novel self-supervised or multimodal architectures such as CLIP, MAE, or DINO to non-standard imaging data including IR, SAR, or hyperspectral data.
- Advanced PyTorch engineering skills and strong mathematical intuition for representation learning.
- Knowledge of C++, Rust, or Go and resource optimization for edge computing.
- Ability to work across research, engineering, hardware, and algorithm teams in support of Harmattan AI’s defense mission.
Categories