Machine Learning Operations Engineer

Job Overview

The Machine Learning Operations Engineer is responsible for development, deployment, and monitoring of machine learning models. The engineer works closely with Data Analysts, ML Engineers in Analytics Exploration and collaborates with Software and Data Engineers to ensure infrastructure and data pipelines are structured to deploy machine learning solutions.    

Job Responsibilities:

• Understands and translates business and functional needs into machine learning problem statements 
• Develops scalable solutions that leverage machine learning models to meet enterprise requirements  
• Works closely with Data Analysts and ML Engineers in Analytics Exploration to develop production grade machine learning models 
• Support life cycle management of deployed ML models (e.g., retraining, new releases, change management, monitoring and troubleshooting). 
• Design, build, automate, deploy, monitor, maintain and own machine learning production model and CI/CD pipelines. 
• Collaborates with development teams to test and deploy machine learning models  
• Designs and implements metrics to continuously evaluate the performance of machine learning solutions 
• Maintains and improves the performance of existing machine learning solutions 
• Keeps abreast with new tools, algorithms and techniques in machine learning and works to implement them in the organization 
• Performs other duties as assigned. 

Qualifications

• More than two years of experience in developing and deploying enterprise-scale machine learning solutions. 
• A minimum of a bachelor's degree in software engineering, data science or computer science or a related quantitative field is required. An advanced degree in a related quantitative field is desired.
• Strong understanding of probability and statistical models with proficiency in machine learning algorithms.
• Strong programming skills with Python and SQL are required.
• Hands-on experience in developing and maintaining APIs (e.g. REST, gRPC, or SOAP) is a must have. 
• Hands-on experience in specifying infrastructure and infrastructure as a code (e.g.: Docker Swarm, CloudFormation) is a must have. 
• Hands-on experience with the concepts and capabilities of a cloud data platform and serverless framework is required.
• Experience in deploying and maintaining reproducible ML models and pipelines.
• Proficient in developing CI/CD pipelines for both code and ML model development, deploying codes/models to production, monitoring models in production, and managing the model lifecycle in a regulated environment.
• Familiar with two or more machine learning frameworks (Sklearn, Tensorflow, PyTorch, etc.).
• Hands-on experience with one or more ML tools like Stagemaker Model Registry, Kubeflow, Metaflow, DVC or MLFlow for lifecycle management or Grafana for monitoring is an asset.
• Proficiency with AWS MLOps technology stack, particularly Feature Store, Model Registry, Sagemaker, CodeBuild, CodePipeline, Lambda, Cloudformation, ECR and Fargate is an asset.
• Practical knowledge of ensemble ML models and explainability techniques (e.g. LIME, SHAP or PDPs) is an asset.
• Ability to effectively communicate technical concepts and results to technical and business audiences in a comprehensive manner.
• Ability to collaborate effectively across multiple teams and stakeholders, including analytics teams, development teams, and operations.
• Required to be highly creative and collaborative.
• Has good judgment, a sense of urgency.
• Commitment to high standards of ethics, regulatory compliance, customer service and business integrity.