Back to results

Stagiaire de recherche - Sécurité de l'IA / Research Intern - AI Safety

  • Montreal, QC, Canada

Temporary

Applications open in a new tab.

Job details

Job type
Temporary
Apply by
Posting language
English · French
Posted

Original job description

English section of this bilingual listing.

Benefits

About Mila

Founded by Professor Yoshua Bengio of the Université de Montréal, Mila brings together researchers specializing in artificial intelligence, and more specifically in machine learning, deep learning and reinforcement learning. Recognized worldwide for its important contributions to the field of deep learning, Mila has particularly distinguished itself in language modeling, machine translation, object recognition and generative models. Since 2017, Mila has been the fruit of a collaboration between Université de Montréal and McGill University, with close links to Polytechnique Montréal and HEC Montréal.

Mila's mission is to be a global hub of scientific advances that inspires innovation and the rise of artificial intelligence (AI) for the benefit of all.

Job Description

We are looking for a talented and passionate applied AI Safety Researcher to join our Mila AI Safety Studio team for a 3 months mandate.

Your role will be to increase the impact of our cutting-edge machine learning research by collaborating closely with researchers and the Mila AI Safety Studio. You will design, implement safety guardrails in the public safety domain, including dynamic benchmarking, red teaming simulation, safety guardrails for harm detection and automatic evaluation. You will thus work in collaboration with our team responsible for the Mila compute cluster to efficiently utilize our computing infrastructure. The project, spanning various challenging scientific domains, offers strong publication potential and contributes to the responsible adoption of AI to solve important societal problems.

You will be expected to create rigorous internal technical documentation and encouraged to publish your work contribution in peer reviewed conferences or papers. If you have a solid background in designing and implementing industrial prototypes and an interest in applied AI research, we strongly encourage you to apply!

Key Responsibilities:

  • Document and write a scientific manuscript on the work of the Mila Safety Studio
  • Collaborate with product managers to understand the requirements necessary for experiment design
  • Work with researchers to select, evaluate, and integrate appropriate AI technologies
  • Identify applied research opportunities stemming from prototypes and user feedback
  • Optimize research, training, and model inference processes

Required Skills and Qualifications:

  • AI research experience with multiple publications in major conferences (NeurIPS, ICML, ICLR)
  • Strong scientific curiosity for applied AI safety research
  • A strong ability to communicate complex scientific concepts to people of various levels of scientific knowledge
  • Ability to collaborate effectively with multidisciplinary teams

Desired Skills and Qualifications:

  • Experience in agile software development and product management
Read the full bilingual original
À propos de Mila

Fondé par le professeur Yoshua Bengio de l'Université de Montréal, Mila rassemble des chercheurs spécialisés en intelligence artificielle et plus précisément en apprentissage automatique, apprentissage profond et apprentissage par renforcement. Reconnu mondialement pour ses importantes contributions au domaine de l'apprentissage profond, Mila s'est particulièrement distingué dans la modélisation du langage, la traduction automatique, la reconnaissance d'objets et les modèles génératifs. Depuis 2017, Mila est le fruit d'une collaboration entre l'Université de Montréal et l'Université McGill, en lien étroit avec Polytechnique Montréal et HEC Montréal.

Mila s'est donné pour mission d'être un pôle mondial d'avancées scientifiques qui inspire l'innovation et l'essor de l'intelligence artificielle (IA) au bénéfice de tous.

Pour en connaitre davantage, veuillez consulter https://mila.quebec/

Description De Poste

Nous recherchons un.e chercheur.euse talentueux.euse et passionné(e) en sécurité de l'IA appliquée pour rejoindre notre équipe Mila AI Safety Studio pour un mandat de 3 mois.

Votre rôle sera d'accroître l'impact de nos recherches de pointe en apprentissage automatique en collaborant étroitement avec les chercheurs et le Mila AI Safety Studio. Vous concevrez et mettrez en œuvre des garde-fous de sécurité dans le domaine de la sécurité publique, incluant le benchmarking dynamique, la simulation de red teaming, les garde-fous de sécurité pour la détection des préjudices et l'évaluation automatique. Vous travaillerez ainsi en collaboration avec notre équipe responsable du cluster de calcul Mila afin d'utiliser efficacement notre infrastructure informatique. Le projet, qui couvre plusieurs domaines scientifiques stimulants, offre un fort potentiel de publication et contribue à l'adoption responsable de l'IA pour résoudre d'importants problèmes sociétaux.

Vous devrez créer une documentation technique interne rigoureuse et vous serez encouragé(e) à publier vos contributions dans des conférences ou des revues révisées par des pairs. Si vous avez une solide expérience dans la conception et la mise en œuvre de prototypes industriels et un intérêt pour la recherche appliquée en sécurité de l'IA, nous vous encourageons vivement à postuler !

Responsabilités clés :

  • Documenter et rédiger un manuscrit scientifique sur le travail du Mila Safety Studio
  • Collaborer avec les gestionnaires de produits pour comprendre les exigences nécessaires à la conception des expériences
  • Travailler avec les chercheurs pour sélectionner, évaluer et intégrer les technologies d'IA appropriées
  • Identifier les opportunités de recherche appliquée découlant des prototypes et des retours d'information des utilisateurs
  • Optimiser les processus de recherche, d'entraînement et d'inférence des modèles

Requirements

Compétences et qualifications requises :

  • Expérience en recherche en IA avec plusieurs publications dans des conférences majeures (NeurIPS, ICML, ICLR)
  • Forte curiosité scientifique pour la recherche appliquée en sécurité de l'IA
  • Grande capacité à communiquer des concepts scientifiques complexes à des personnes de divers niveaux de connaissances scientifiques
  • Capacité à collaborer efficacement avec des équipes multidisciplinaires

Compétences et qualifications souhaitées :

  • Expérience en développement logiciel agile et en gestion de produits
  • Doctorat (PhD) en informatique, science des données ou domaines pertinents
  • Excellente maîtrise du langage de programmation Python

Benefits

About Mila

Founded by Professor Yoshua Bengio of the Université de Montréal, Mila brings together researchers specializing in artificial intelligence, and more specifically in machine learning, deep learning and reinforcement learning. Recognized worldwide for its important contributions to the field of deep learning, Mila has particularly distinguished itself in language modeling, machine translation, object recognition and generative models. Since 2017, Mila has been the fruit of a collaboration between Université de Montréal and McGill University, with close links to Polytechnique Montréal and HEC Montréal.

Mila's mission is to be a global hub of scientific advances that inspires innovation and the rise of artificial intelligence (AI) for the benefit of all.

To find out more, please visit https://mila.quebec/

Job Description

We are looking for a talented and passionate applied AI Safety Researcher to join our Mila AI Safety Studio team for a 3 months mandate.

Your role will be to increase the impact of our cutting-edge machine learning research by collaborating closely with researchers and the Mila AI Safety Studio. You will design, implement safety guardrails in the public safety domain, including dynamic benchmarking, red teaming simulation, safety guardrails for harm detection and automatic evaluation. You will thus work in collaboration with our team responsible for the Mila compute cluster to efficiently utilize our computing infrastructure. The project, spanning various challenging scientific domains, offers strong publication potential and contributes to the responsible adoption of AI to solve important societal problems.

You will be expected to create rigorous internal technical documentation and encouraged to publish your work contribution in peer reviewed conferences or papers. If you have a solid background in designing and implementing industrial prototypes and an interest in applied AI research, we strongly encourage you to apply!

Key Responsibilities:

  • Document and write a scientific manuscript on the work of the Mila Safety Studio
  • Collaborate with product managers to understand the requirements necessary for experiment design
  • Work with researchers to select, evaluate, and integrate appropriate AI technologies
  • Identify applied research opportunities stemming from prototypes and user feedback
  • Optimize research, training, and model inference processes

Required Skills and Qualifications:

  • AI research experience with multiple publications in major conferences (NeurIPS, ICML, ICLR)
  • Strong scientific curiosity for applied AI safety research
  • A strong ability to communicate complex scientific concepts to people of various levels of scientific knowledge
  • Ability to collaborate effectively with multidisciplinary teams

Desired Skills and Qualifications:

  • Experience in agile software development and product management
  • PhD in Computer Science, Data Science or relevant domains
  • Excellent proficiency in the Python programming language
More about this job

Responsibilities

Design and implement AI safety guardrails for public safety, including dynamic benchmarking, red-team simulations, harm detection, and automated evaluation, while making effective use of Mila’s computing infrastructure. Document the work and prepare a scientific manuscript, collaborate with product managers and researchers on experiments and AI technology, identify research opportunities from prototypes and user feedback, and optimize research, training, and inference processes.

Requirements

Applicants must have AI research experience with multiple publications at major conferences such as NeurIPS, ICML, or ICLR, strong interest in applied AI safety, effective communication skills, and the ability to collaborate across disciplines. Desired qualifications include a PhD in computer science, data science, or a related field, experience with agile software development and product management, and excellent Python proficiency.

Skills

  • AI Safety Research
  • Machine Learning
  • AI Guardrails
  • Dynamic Benchmarking
  • Red Teaming
  • Harm Detection
  • Automated Evaluation
  • Scientific Writing
  • Experiment Design
  • AI Technology Evaluation
  • Applied Research
  • Model Training
  • Model Inference Optimization
  • Python
  • Agile Software Development
  • Product Management

Visa sponsorship

Not detected in the job text

Categories

  • Science & Research
  • Technology
  • Software
  • Security & Safety
  • Data & Analytics

Keywords

  • Artificial Intelligence
  • AI Safety
  • Machine Learning
  • Deep Learning
  • Reinforcement Learning
  • Public Safety
  • Safety Guardrails
  • Dynamic Benchmarking
  • Red-Team Simulation
  • Harm Detection
  • Automated Evaluation
  • AI Research
  • Applied Research
  • Scientific Manuscript
  • Technical Documentation
  • Experiment Design
  • AI Technology Integration
  • User Feedback
  • Computing Infrastructure
  • Model Training
  • Model Inference
  • Python
  • Agile Software Development
  • Product Management
  • NeurIPS
  • ICML
  • ICLR

Report this job