Applied Machine Learning Scientist I, Tabular Deep Learning Model Validation

Work Location:

Toronto, Ontario, Canada

Hours:

37.5

Line of Business:

Analytics, Insights, & Artificial Intelligence

Pay Details:

$105,500 - $125,000 CAD

The pay details posted reflect a temporary market premium specific to this role that is reassessed annually.

TD is committed to providing fair and equitable compensation opportunities to all colleagues. Growth opportunities and skill development are defining features of the colleague experience at TD. Our compensation policies and practices have been designed to allow colleagues to progress through the salary range over time as they progress in their role. The base pay actually offered may vary based upon the candidate's skills and experience, job-related knowledge, geographic location, and other specific business and organizational needs.

As a candidate, you are encouraged to ask compensation related questions and have an open dialogue with your recruiter who can provide you more specific details for this role.

Job Description:

*Department Overview

TD Model Validation (MV) group is responsible for the independent validation and approval of models used for Generative AI, Natural Language Processing (NLP), Tabular Deep Learning, credit, fraud, and marketing models. The Artificial Intelligence/Machine Learning (AI/ML) MV team is responsible for the validation of all AI/ML models used across the Bank for various use cases.

*Job Description

The position reports to Senior Applied Machine Learning Scientist in the AI/ML Model Validation team and is primarily focused on the validation and review of Deep Learning models.

Detailed accountabilities include:

• Research, validate, and apply new techniques in testing deep learning models applied to tabular/structured datasets.
• Stay up to date with advancements in the field of AI including major publications, foundation models, evaluation metrics, explainability, technology stacks, and datasets.
• In depth knowledge of techniques and developments in the field of AI/ML and share knowledge with business partners and senior management.
• Develop/implement AI/ML model validation methodologies and standards, especially for Tabular Deep Learning / Foundation models. Ensure that the validation methodologies and standards are in line with industry best practices and address regulatory and audit requirements.
• Work in Distributed Cluster / Cloud environments with large-scale datasets ranging from transactions to large document collections.
• Develop and apply a variety of statistical tests and modeling techniques to identify/recommend improvements to models and undertake related initiatives. Implement benchmark models as applicable.
• Conduct R&D in the area of Tabular Deep Learning and Foundation Models evaluation, testing, explainability.
• Communicate findings and recommendations to both technical and non-technical stakeholders.

The position involves working effectively with different internal partners such as AI2, Layer6, P&T, and FCRM.

*Job Requirements

• Strong quantitative skills with an advanced degree in one or more of the following areas: computer science, machine learning / AI, engineering, statistics, mathematics, or physics.
• Experience with and strong knowledge of AI/ML methodologies including Deep Learning, modern Natural Language Processing (NLP), Transformers, Diffusion models, and Bagging/Boosting methods.
• Experience with Deep Learning technology stacks and libraries such as PyTorch, Tensorflow/Keras etc.
• Stay up to date with the latest advancements in Tabular Deep Learning / Foundation models, Machine Learning, and Cloud technologies.
• Proficient in one or more scripting/programming languages such as Python, Java, Scala, Pyspark.
• Familiarity with cloud platforms (e.g., Azure Databricks, Azure ML Studio).
• Familiarity with Data Structures, Algorithm design, and principles of Object-Oriented Programming (OOP).
• Knowledge of machine learning explain-ability/interpretability algorithms.
• Excellent verbal and written communication skills. The position requires writing clean technical reports.
• Ability to work independently and collaboratively in a fast-paced, dynamic environment. Great time management and multitasking skills with minimal supervision.

* Preferred Qualifications

• Publications in the relevant conference and journals are a plus.
• Ability to implement AI/ML algorithms from academic research papers is a plus.

*Aperçu du service

Le groupe Validation des modèles TD est responsable de la validation et de l’approbation indépendantes des modèles utilisés pour l’intelligence artificielle générative, le traitement du langage naturel, l’apprentissage profond pour les données tabulaires, le crédit, la gestion de la fraude et le marketing. L’équipe Validation des modèles, Intelligence artificielle (IA) et Apprentissage automatique (AA) est responsable de la validation de tous les modèles d’intelligence artificielle et d’apprentissage automatique utilisés à l’échelle de la Banque pour divers cas d’utilisation.

*Description du poste

Le poste relève du scientifique principal, Apprentissage automatique appliqué au sein de l’équipe Validation des modèles, IA et AA et est principalement axé sur la validation et l’examen des modèles d’apprentissage profond.

Voici les responsabilités détaillées du poste :

• Rechercher, valider et appliquer de nouvelles techniques de tests de modèles d’apprentissage profond appliqués aux ensembles de données tabulaires/structurés.
• Se tenir au courant des progrès dans le domaine de l’IA, y compris les principales publications, les modèles de fondation, les mesures d’évaluation, l’explicabilité, les piles technologiques et les ensembles de données.
• Maintenir une connaissance approfondie des techniques et des développements relatifs au secteur de l’IA et l’AA et partager ses connaissances avec les partenaires d’affaires et la haute direction.
• Élaborer et mettre en œuvre des méthodes et des normes de validation des modèles d’IA et d’AA, en particulier pour les modèles de fondation et d’apprentissage profond pour les données tabulaires. Veiller à ce que les méthodes et les normes de validation soient conformes aux pratiques gagnantes du secteur et à ce que les exigences en matière de réglementation et d’audit soient traitées.
• Travailler dans des environnements infonuagiques ou une grappe de serveurs distribués avec des ensembles de données à grande échelle, allant des opérations aux grandes collections de documents.
• Élaborer et appliquer une variété de tests statistiques et de techniques de modélisation pour cerner/recommander des améliorations à apporter aux modèles et entreprendre des initiatives connexes. Mettre en œuvre des modèles de référence, le cas échéant.
• Effectuer la recherche et le développement liés à l’évaluation, à la mise à l’essai et à l’explicabilité des modèles de fondation et d’apprentissage profond pour les données tabulaires.
• Communiquer les constatations et les recommandations aux intervenants techniques et non techniques.

Le titulaire du poste doit travailler efficacement avec différents partenaires internes, comme les équipes Analyses, Idées et Intelligence artificielle, Layer 6, Plateformes et Technologie et Gestion des risques liés aux crimes financiers.

*Exigences du poste

• Solides compétences en analyse quantitative et diplôme d’un cycle supérieur en informatique, en apprentissage automatique, en IA, en génie, en statistique, en mathématique ou en physique.
• Solides connaissances des méthodologies relatives à l’IA et l’AA (et expérience connexe), y compris l’apprentissage profond, le traitement du langage naturel moderne, les transformateurs, les modèles de diffusion et les méthodes de bagging et d’optimisation.
• Expérience des bibliothèques et des piles technologiques d’apprentissage profond, comme PyTorch et Tensorflow/Keras.
• Connaissances à jour des dernières avancées liées aux modèles de fondation et d’apprentissage profond pour les données tabulaires, à l’AA et aux technologies infonuagiques.
• Maîtrise d’un ou de plusieurs langages de script/programmation, comme Python, Java, Scala ou Pyspark.
• Connaissance des plateformes infonuagiques (p. ex. Azure DataBricks, studio d’apprentissage automatique d’Azure).
• Connaissance des structures de données, de la conception d’algorithmes et des principes de la programmation orientée objet.
• Connaissance des algorithmes d’interprétation/d’explicabilité de la capacité d’apprentissage automatique.
• Excellentes aptitudes pour la communication orale et écrite. La personne qui occupera le poste devra rédiger des rapports techniques clairs.
• Capacité à travailler de façon autonome et collaborative dans un environnement dynamique au rythme trépidant. Excellentes compétences en gestion du temps et capacités à gérer plusieurs tâches en même temps avec un minimum de supervision.

*Compétences recherchées

• Actes de colloque et publications scientifiques pertinents, un atout.
• Capacité à mettre en œuvre des algorithmes d’intelligence artificielle et d’apprentissage automatique à partir d’articles de recherches universitaires, un atout.

Who We Are:

TD is one of the world's leading global financial institutions and is the fifth largest bank in North America by branches/stores. Every day, we strive to make every interaction, product, and experience remarkably human and refreshingly simple for over 27 million households and businesses in Canada, the United States and around the world. More than 95,000 TD colleagues bring their skills, talent, and creativity to foster deeper relationships, ensure disciplined execution, and build a simpler, faster banking experience. TD is deeply committed to being a leader in client experience, that is why we believe that all colleagues, no matter where they work, are client facing. Together, we are reimagining what banking can be for our clients, colleagues and communities.

Our Total Rewards Package
Our Total Rewards package reflects the investments we make in our colleagues to help them and their families achieve their financial, physical, and mental well-being goals. Total Rewards at TD includes a base salary, variable compensation, and several other key plans such as health and well-being benefits, savings and retirement programs, paid time off, banking benefits and discounts, career development, and reward and recognition programs. Learn more

Additional Information:
We’re delighted that you’re considering building a career with TD. Through regular development conversations, training programs, and a competitive benefits plan, we’re committed to providing the support our colleagues need to thrive both at work and at home.

Please be advised that this job opportunity is subject to provincial regulation for employment purposes. It is imperative to acknowledge that each province or territory within the jurisdiction of Canada may have its own set of regulations, requirements.

Colleague Development

If you’re interested in a specific career path or are looking to build certain skills, we want to help you succeed. You’ll have regular career, development, and performance conversations with your manager, as well as access to an online learning platform and a variety of mentoring programs to help you unlock future opportunities.

If you’re passionate about helping clients and building deep, lasting relationships, TD offers diverse career paths where you can grow your expertise and make a meaningful impact.  

We're committed to your success and foster a respectful workplace where diverse perspectives are valued, everyone has fair opportunities to grow, and you can unlock your full potential to achieve your career goals. Here at TD, we hire and develop the best.

Training & Onboarding
We will provide training and onboarding sessions to ensure that you’ve got everything you need to succeed in your new role.

Interview Process 
We’ll reach out to candidates of interest to schedule an interview. We do our best to communicate outcomes to all applicants by email or phone call.

Accommodation
Your accessibility is important to us. Please let us know if you’d like accommodations (including accessible meeting rooms, captioning for virtual interviews, etc.) to help us remove barriers so that you can participate throughout the interview process.

We look forward to hearing from you!

Language Requirement (Quebec only):

Sans Objet