Companies in this catalog do not charge applicants. Never send money to apply.
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.
Programming: Strong proficiency in Python and familiarity with TypeScript/Node.js for deployment integration.
Methodology: Deep understanding of MLOps best practices, responsible AI principles, and monitoring concepts.
AI Agent Focus You will ensure the reliability and continuous improvement of the core AI layer.
LLM Operationalization: Implement specific pipelines for the fine-tuning, validation, and deployment of Large Language Models (LLMs) used in our AI agents.
Agent Performance Tracking: Develop metrics and tracking systems to measure the business impact and operational efficiency of multi-agent systems and recommendation engines.
Framework Integration: Operationalize models built using frameworks like LangChain or LlamaIndex, ensuring they are secure, versioned, and scalable in a production environment.
Success Metrics & Career Path
Performance will be measured by:
Deployment Velocity: Speed and reliability of deploying new or retrained models to production.
Model Performance: Maintaining model accuracy and minimizing performance drift in production.
Pipeline Automation: Percentage of the ML lifecycle (training, validation, deployment) that is fully automated.
Mentorship Structure: Reports to the Solution Architect or Head of Technology, collaborating closely with Data Architects, Data Scientists, and SREs to maintain a reliable AI ecosystem.
