Deep Learning Engineer Intern

At Hadō, we are on a mission to restore trust in the digital world.

In an era of rampant misinformation and Generative AI, ensuring the authenticity and traceability of digital content has become a critical challenge for society. Hadō develops an invisible watermarking technology powered by deep learning that enables creators, brands, and industries to track their assets, prevent leaks, and verify the origin of content.

We are an early-stage startup, incubated at Agoranov (the home of Dataiku, Doctolib, Criteo) & WILCO, one of Europe's best startup accelerators. We are members of the C2PA standard body as well.

We are looking for a Deep Learning Engineer Intern to join our R&D team. You will work hand-in-hand with Bruno Spilak (CTO, PhD in Machine Learning) to solve one of the hardest challenges in watermarking: building the next generation of deep learning models for universal and indestructible digital provenance.

Your Role & Responsibilities

You will dive deep into the core technology of Hadō. Your main objective will be to strengthen our deep learning architectures to ensure the universal durability of our watermarks across any digital lifecycle. You will work closely with Bruno (CTO) on the following missions:

• Design End-to-End Robust Architectures:
  Develop and train encoder-decoder models capable of embedding high-capacity data while remaining perfectly imperceptible to the human eye.
• Build an Adversarial Augmentation Pipeline:
  Create a sophisticated "distortion layer" in our training loop to simulate a wide array of attacks (from standard geometric transforms to neural-based modifications) reducing the need for manual dataset curation.
• Optimize for Scalability & Latency:
  Assist the CTO in refining our models for production, ensuring they can process high-resolution documents.
• Tech Watch & Innovation:
  Review state-of-the-art papers in Deep Steganography, Computer Vision, and AI Security to identify and test new approaches that could become part of Hadō's tech.

Our Research Stack

• Languages:
  Python (Proficient).
• ML Frameworks:
  PyTorch (mandatory), HuggingFace.
• Environment:
  AWS EC2 and Sagemaker
• Tracking:
  MLflow / TensorBoard.

Who You Are

We are looking for a rigorous, scientifically-minded individual who enjoys solving complex mathematical and algorithmic problems and turning ideas into working models.

• Education:
  You are in your final year of a Master's degree or Engineering School (M2 / End-of-studies internship) with a specialization in Machine Learning, Computer Vision, or Applied Mathematics. Prior research experience (projects, publications, competitions) is a strong plus.
• Technical Proficiency:
  You have strong coding skills in Python and practical experience with PyTorch. You understand the mechanics of geometric transformations and signal degradation.
• Deep Tech Appetite: You are interested in the physics of image formation (how a camera sees a screen or paper) and how to model it with Deep Learning.
• Scientific Method:
  You are rigorous in your experimental protocols, documenting your hypotheses, results, and failures clearly.
• Team Spirit:
  You value transparency, collaboration and are eager to work in a small, high-performing team.

What We Offer

• Best Mentorship:
  Work directly with a PhD CTO who will guide you through advanced research methodologies.
• Real Impact:
  Your code will directly contribute to securing the world's digital information by making content provenance indestructible and universally verifiable.
• Hiring Potential:
  We are looking for talent to stay A full-time position (CDI) is the target outcome of this internship.
• Top-Tier Ecosystem:
  Work from our office at Agoranov in Paris and be part of WILCO, one of Europe's best startup accelerators. You will be surrounded by the best French deep tech entrepreneurs.
• A Cool & Ambitious Partnership:
  We work hard to build a world-class product, but we are also a fun-loving team that values true partnership and knows how to celebrate wins together.

The Interview Process

1. Intro Call with one of the founders to discuss the vision and fit (30 min).
2. Technical Interview with Bruno (CTO): Discussion on your previous ML projects, math basics, and paper reading (1h).
3. Meet the Team: Lunch or coffee at Agoranov to ensure cultural fit.

Ready to build the invisible shield for physical and digital content?

Apply now