Staff MLOps Developer, Risk

In this role, you will have the opportunity to:

• Automated Model Recalibration Pipeline: Implement and deploy an automated pipeline for model recalibration and re-fitting across key models, starting with those for fraud detection. Your work will ensure our models maintain optimal performance through systematic refreshing as market conditions evolve and threat actors' tactics change.
• Wealthsimple AI Foundation Model: Contribute to the continued development and productionization of our foundational AI model. This project requires both data science expertise and engineering rigour to transform how we understand our clients. You will apply advanced machine learning techniques while building scalable training pipelines and optimizing model performance. Your work will involve designing systems that effectively process diverse client signals (trades, transfers, clicks, purchases) into a unified high-dimensional embedding, then engineering these solutions from development to production. This versatile embedding will serve as the core intelligence layer for numerous downstream applications, including:
• Fraud Detection: classify signals associated with known fraud types.
• Client Segmentation: group clients based on their unique behaviours.
• Outlier Detection: identify abnormal behaviours and patterns beyond traditional fraud, enhancing our ability to maintain a secure and trustworthy platform.
• Model Explainability:  Drive model explainability (xAI) initiatives to ensure ML models meet regulatory requirements and can be successfully deployed to production environments. Develop and implement interpretability techniques that make complex models transparent to stakeholders, satisfying compliance standards while enabling faster model approval and deployment cycles.
• MLOps Security & Infrastructure Expertise: Lead the implementation of best practices in our machine learning operations to create a secure, scalable infrastructure. You will architect robust ML systems that protect against cybersecurity vulnerabilities including data poisoning, model evasion, and adversarial attacks, while ensuring our decision infrastructure remains performant and maintainable at scale.

What we are looking for:

• 5+ years of experience as a data scientist and/or machine learning engineer, with a focus on advanced models in finance.
• Expertise in Python, with a strong familiarity with key data science & machine learning libraries and tools. (PyTorch/TensorFlow/LightGBM/Scikit-Learn).
• Ownership of the design and development of an ML platform: demonstrated experience in productionalizing ML models, with a track record of owning the platform for ML model development, from prototyping to deployment and maintenance.
• Solid knowledge of data engineering and preprocessing, including handling large datasets, ensuring data quality, and managing data pipelines.
• Excellent communication skills, with the ability to convey complex ideas in a simple, clear, and engaging manner for a variety of audiences.
• Outstanding interpersonal skills, demonstrated through the ability to thrive in highly collaborative settings and contribute positively to team-oriented environments.
• Familiarity with key deep learning and LLM concepts such as tokenization, fine-tuning, sequence prediction, embedding.
• Experience optimizing the runtime of training or inference on GPUs including parallel training strategies, optimal batching, mixed precision.
• Familiarity with AI ethics and responsible AI practices, with a commitment to ensuring that our solutions are not only effective but also ethical and fair.
• Experience with relevant data platforms like kafka, snowflake, redshift.
• Experience with ML development platform and cloud environments like AWS and SageMaker.
• Experience with distributed systems, containerization (Docker/Kubernetes), and workflow orchestration (Airflow).
• Experience with CI/CD tools and working with Git.