MLOps Engineer Internship

About this role

Overview: Operationalizing AI and Personalization The MLOps Engineer is crucial for bridging the gap between data science and production, responsible for the reliable, scalable, and secure deployment of machine learning models. You will operationalize the models powering our AI agents and the e-commerce personalization systems, ensuring continuous integration, delivery, and monitoring of our predictive analytics and recommendation engines.

Internship Details Duration: 3 months Start Date: Immediate Location: Remote Stipend: None initially. Based on your first-quarter performance, you may be offered a paid full-time opportunity, or even be absorbed directly by the client as an FTE. Key Responsibilities & Core Projects You will build the automation infrastructure that turns static models into continuously improving production systems. Model CI/CD Pipelines: Design and build robust CI/CD pipelines dedicated to the machine learning lifecycle: automated model training, validation, and deployment using tools integrated with our main Makefile CI/CD setup.

Model Monitoring & Tracking: Implement comprehensive monitoring and alerting for model performance (e.g., drift detection, prediction accuracy, latency) and track experiments and model artifacts using version control tools. Production Deployment: Operationalize the deployment of ML models powering AI agents and e-commerce services, ensuring they integrate seamlessly into the NestJS modular monolith architecture. Versioning & Feature Stores: Manage model versioning and lineage.

Collaborate on the design and maintenance of a centralized Feature Store to ensure consistent data for training and serving. Experimentation Infrastructure: Implement and manage the infrastructure necessary for A/B testing different model versions or personalization strategies in a production environment (e-commerce storefront). Retraining Workflows: Define and automate the model retraining workflows based on data drift or performance degradation triggers, ensuring models remain relevant to the dynamic supply chain and customer behavior.

Required Technologies & Tools Candidates must possess hands-on expertise in the tools and methodologies used for production ML and MLOps: MLOps Tools: Experience with model registries, experiment tracking, and serving platforms (e.g., MLflow, Kubeflow, Sagemaker). CI/CD & Automation: Proficiency in building pipelines (using Python/Bash scripting) and experience with Docker and Terraform. Data & Compute: Experience managing data pipelines for ML (ETL/ELT) and optimizing compute resources for training and inference.