Data Engineer (Detect & Track Distillation)

Harmattan AI builds autonomous, scalable defense systems for allied forces and has recently closed a $200 million Series B round that values the company at $1.4 billion. The Data Engineer role will own the data layer that supplies the machine‑learning teams, turning raw field logs and public sources into clean, versioned datasets ready for training. Day‑to‑day work includes ingesting and decoding unstructured sensor streams such as video, aligning multiple modalities in time and space, and curating the data by parsing, filtering and de‑duplicating. You will define labeling needs, manage the end‑to‑end dataset construction workflow, keep versioning and lineage information, store data in efficient training‑ready formats, and ensure fast delivery of documented datasets so model training is not I/O‑bound. The position requires a STEM degree or equivalent experience, proven experience building large‑scale pipelines for unstructured data, strong Python skills, and familiarity with optimizing data loaders for frameworks like PyTorch. Desirable background includes work with multimodal sensor data, labeling pipelines, dataset‑versioning tools, and distributed processing. The role is based in Paris, Lausanne or Zurich, offering early‑stage influence on how data fuels deep‑learning pipelines within a mission‑driven, high‑rigor environment.