Senior Machine Learning Scientist - Foundation Models

PassFort; Paris; Hybride; EUR 90 000 - 130 000; Plein temps

Avantages offerts par ce poste

Remote-friendly schedule

Office in Paris

Budget for remote work equipment

Gymlib subscription

Meal card (France)

Résumé du poste

Spore.Labs is seeking a senior ML scientist to co-build a foundation model mapping multimodal signals to biologically grounded embeddings. You will collaborate with microbiologists and optical physicists to ensure interpretability and publish results.

You will prototype scalable training pipelines on multi-GPU cloud infrastructures, reproduce state-of-the-art results, and contribute to IP through publications and patents. Fluency in English is expected.

Qualifications

• Academic background: engineer degree, Master's, or PhD in ML with publications in top ML conferences or journals.
• Professional experience: minimum of 4 years hands-on ML experience, incl. at least one large-scale training project.
• Programming: proficient in Python and PyTorch; comfortable with ML ops; CI practices.
• ML knowledge: strong theory and causality skills.
• Reliability and autonomy: able to organize work independently; manage compute time and data availability.
• Team mindset: collaborate across tasks; strong scientific English communication.

Responsabilités

• Coordinate with ML engineers to build scalable foundation model training pipelines.
• Benchmark AMR prediction approaches and ensure reproducibility.
• Develop and implement deep learning models for multimodal data (transformers, GNNs, autoencoders).

Connaissances

Multimodal ML

Foundation models

Python

PyTorch

Distributed training

Biology knowledge

Formation

PhD in ML/AI

Publications in ML venues

Outils

PyTorch

Cloud platforms

CI/CD

Description du poste

About Us

Spore.Bio is a deeptech startup founded in 2023 that is redefining microbiological quality control in pharmaceutical, food & beverage, and cosmetics manufacturing. Where Spore.Bio deploys biophotonic and deep-learning technology on factory floors, Spore.Labs takes Spore.Bio's core technology into new territory: from AMR detection to microbiome research and beyond.

Spore.Labs is looking for a new talent ready to take real ownership, prototype solutions, build incrementally a robust ML pipeline to train a multimodal foundation model, benchmark SOTA approaches and grow alongside the department they help build.

About the Role

Data-driven AMR prediction requires access to signals that reveal both the root causes of resistance and the downstream consequences that expose activated pathways. The senior ML scientist will co-build a foundation model mapping multiple modalities (whole genome sequencing, transcriptomics, and proteomics, spectroscopy data, multi-spectral images) to biologically grounded embeddings capturing the state of microorganisms.

Working closely with the Spore.Labs ML Coordinator, you will evaluate multimodal integration architectures and propose original approaches that incentivise the model to learn causal relationships aligned with biomolecular rules, leveraging interventional data. You will also reproduce state‑of‑the‑art results from the literature and contribute to model validation campaigns.

Missions

• Coordinate with ML engineers to contribute in building a foundation model training pipeline at scale on multi‑GPU cloud infrastructures.

• Benchmark AMR prediction approaches from the literature and ensure reproducibility of the results

• Develop and implement original deep learning models (transformers, graph neural networks, autoencoders) for representation learning from multiple modalities.

• Maintain an up‑to‑date knowledge of recent foundation / world model literature

• Collaborate closely with microbiologists, omcis experts, computer vision experts, and optical physicists to ensure biological interpretability of AI models.

• Contribute to scientific publications, conferences, and intellectual property (patents) highlighting novel insights from the trained foundation model.

About You

We value curiosity, initiative, and a willingness to learn. Not every box needs to be ticked to apply; we are recruiting for trajectory as much as for current skills. Strong motivation to work at the interface of biology, optics, and AI in a multidisciplinary deeptech environment is a genuine plus.

• Academic background: engineer degree, Master's degree, or PhD in machine learning. Proven record of publications in top tier ML conferences (NeurIPS, ICML, ICLR, AISTATS, CVPR) or journals (Nature journals, JMLR, IEEE transactions, or equivalent).

• Professional experience: A minimum of 4 years of proven hands‑on ML experience, including at least one large‑scale training project. Previous experience in multimodal architectures is a strong plus.

• Programming: Proficient in python and torch, comfortable with ML ops stack, documents systematically, uses continuous integration tools and best coding practices.

• Machine learning: Excellent command of ML theory and model classes, strong statistical and causality skills

• Reliability and autonomy: Able to organise their working day independently once roadmaps are established, anticipate scheduling constraints (computation time, data availability), check for spurious correlation and deliver statistically meaningful conclusions.

• Team mindset: Comfortable contributing across tasks in a small laboratory environment. Excellent written and verbal scientific communication in English.

Why Join Us?

• Be part of a core interdisciplinary team driving the development of a disruptive diagnostic technology with real global health impact
• Work at the frontier of microbiology, advanced optics, and machine learning in a team where scientific rigour is the default
• Process real clinical and research samples that feed a concrete diagnostic platform from an early career stage
• Genuine room to grow: as the laboratory scales, responsibility and independence scale with it, in a permanent position (CDI) in Paris

What we offer

We believe that flexibility and trust are important parts of a company. Our work environment reflects this thanks to:

Flexible remote: If you live in Paris, you can work from our office or from your place with no constraints.

On top of that, we offer many perks such as:

• A budget for remote work equipment
• A Gymlib subscription for you to stay in shape wherever you are
• Premium health insurance (Alan in France)
• A Swile card for your meals, if you are based in France
• Frequent team events and in‑person gatherings every quarter!

Recruitment Process

• 30-minute introductory call with a team member
• Home assignment technical test (2 days to handover the work)
• 45-minute technical interview (questions on ML/Stats and scientific practices)
• Lab visit
• Founders Interview