Data Engineer: Build Scalable ML Data Pipelines

Signal Processing Engineer (RF Processing & Analytics); Paris; Sur place; EUR 70 000 - 110 000; Plein temps

Résumé du poste

Harmattan AI, with hubs in Paris, Lausanne, and Zurich, is seeking a Data Engineer to own the data layer feeding ML teams. You will turn raw field logs and public datasets into clean, versioned, training-ready data, enabling modelers to focus on model design and training.

Responsibilities include building ingestion pipelines, multimodal data alignment, data curation and labeling workflows, and maintaining dataset versioning and lineage for reproducibility across training runs.

Qualifications

• Experience building pipelines for unstructured data at scale (video, images, or sensor data).
• Strong in Python and data engineering; optimize data loaders for training frameworks like PyTorch.
• Educational background in a STEM field or equivalent practical experience.

Responsabilités

• Ingestion pipeline: ingest, decode, and store raw, unstructured field data.
• Multimodal alignment: align data streams temporally and spatially for training.
• Curation: transform ingested data and public datasets into high-value, versioned data.
• Data & labeling requirements: define data to gather and labeling workflow.
• Coordinate with teams responsible for data gathering and labeling.
• Dataset construction: build task-specific datasets for training/evaluation.
• Versioning & lineage: maintain dataset versions and lineage for reproducibility.
• Storage & formats: store data in training-ready formats and manage storage tiering.
• Efficient delivery: provide clean datasets and tooling for loading to training pipelines.

Connaissances

Python

Data engineering

PyTorch

Data pipelines

Formation

STEM degree or equivalent experience

Outils

PyTorch

Dataset versioning tooling

Distributed data processing

Description du poste

Harmattan AI, with hubs in Paris, Lausanne, and Zurich, is seeking a Data Engineer to own the data layer feeding ML teams. You will turn raw field logs and public datasets into clean, versioned, training-ready data, enabling modelers to focus on model design and training.

Responsibilities include building ingestion pipelines, multimodal data alignment, data curation and labeling workflows, and maintaining dataset versioning and lineage for reproducibility across training runs.