Machine Learning Engineer – Climate tech

Il y a 2 jours

Paris, Île-de-France CarbonFarm Technology Temps plein 70 000 € - 110 000 € Contrat

What we are looking for

CarbonFarm’s work combines remote sensing data, agricultural data and machine learning. We are looking for a climate-passionate Machine Learning Engineer to turn terabytes of satellite data into useful insights about farming practices.

Data is at the heart of everything we do, and we have a lot of it You’ll work with a wide variety of satellite imagery and Internet-of-things data to build a comprehensive understanding of agricultural practices around the world. Our team and partners are spread across dozens of countries, so you’ll need to be a clear technical communicator who’s capable of understanding different points of view.

Responsibilities

  • Collaborate with fellow engineers (ML, AI, Data) to develop models and algorithms that identify how farmers are cultivating their land from satellite and IoT data
  • Propose and implement new pre-training and fine-tuning approaches for deep models
  • Work with our chief architect to deploy models in production
  • Research different approaches in remote sensing from the scientific community
  • Test and validate models, with a particular focus on temporal/geospatial cross-validation and the mitigation of bias and leakage
  • Leverage AI tooling to drive rapid iteration, with a strong bias for technical accuracy

Requirements

  • Master’s degree or equivalent in data science, computer science, or remote sensing
  • 3+ years of experience with PyTorch (preferred), TensorFlow, or related frameworks
  • Comfortable with a variety of different database technologies (SQL, OLAP)
  • Strong foundations in statistics and probability
  • Flexible and focused on solutions
  • Organised and self-sufficient
  • Fluent in English
  • Paris-based or willing to relocate, with the ability to work in person with the team

Would be nice

  • Familiarity working in an agile environment
  • Experience using cloud-based computing (GCP or similar)
  • Previous exposure to the remote sensing domain