Skip to content
Intesa Sanpaolo loses €95M to a cloned voice.
All open roles
EngineeringFull timeHybrid, MENA

AI engineer I (Detection & optimisation)

You are a high speed executor and builder. Working directly with the Senior Lead, you will help build, test, and refine our classification models from the ground up. If you are obsessed with PyTorch and the challenge of "Real vs. Fake," this is the place to grow.

Key responsibilities

  • Build & experiment: execute the initial codebases for detection in audio, video, and documents using PyTorch.
  • Data curation: assist in building the specialized datasets required to train models to detect sophisticated forgeries and deepfakes.
  • Fine tuning & training: run experiments, track performance metrics, and assist in model quantization and pruning to ensure our engine is fast and lean.

Technical requirements

  • Experience: open to high-potential candidates who can demonstrate a strong portfolio or prior experience in ML development.
  • PyTorch fluent: you can build, debug, and optimise models in PyTorch with minimal supervision.
  • Large Multimodal Models (LMMs): a solid understanding of LMM architectures, with a strong desire to master fine tuning, optimisation, and deployment of models that handle diverse data types (vision, audio, and text).
  • Cloud basics: familiarity with AWS (training) and GCP (deployment) environments.
  • The "detective" mindset: a passion for identifying the subtle nuances of deepfakes and document forgery.

Apply for this role

PDF only, max 10 MB. We use your information solely to evaluate this application.