Cloudera Machine Learning Engineer key terms

A browser-based integrated development environment for data scientists to create projects, write and run code, and manage sessions on a Cloudera cluster.

A container for files, code, and configuration in Cloudera AI that enables collaboration and version control, tied to a Git repository or local storage.

A running compute instance (container) launched from a project with a chosen engine and resource profile, where code executes and files are accessible.

A container image that defines the software environment (e.g., Python, R, Spark) for a session, including libraries and tools.

A predefined set of CPU, memory, and GPU resources assigned to a session or job, determining compute capacity and cost.

The capability to link a Cloudera AI project to a remote Git repository for version control, enabling clone, commit, push, and pull operations.

A stored configuration (e.g., HDFS, S3, JDBC) that allows projects to access external data sources securely without embedding credentials in code.

The process of importing data from various sources into a project or data lake for analysis and modeling.

A distributed computing framework for large-scale data processing, often used in Cloudera AI for data preparation and feature engineering.

A data warehouse infrastructure built on Hadoop that provides SQL-like querying and metadata management for structured data.

An open table format for huge analytic datasets that supports schema evolution, time travel, and ACID transactions on data lakes.

A metadata management system (e.g., Apache Atlas) that tracks data assets, schemas, lineage, and access permissions across the platform.

The record of data origin, movement, and transformation over time, used for auditing, debugging, and impact analysis.

An interactive web-based document that combines code, equations, visualizations, and narrative text, commonly used for experimentation in Cloudera AI.

A tracked run of a model training script with specific parameters, metrics, and artifacts, used for comparison and reproducibility.

An open-source platform for managing the machine learning lifecycle, including tracking experiments, packaging models, and deploying them.

A centralized repository for storing, versioning, and managing machine learning models, including stage transitions (e.g., staging, production).

The process of optimizing model hyperparameters (e.g., learning rate, tree depth) using search algorithms to improve performance.

A centralized system for storing and serving curated features for machine learning, ensuring consistency between training and serving.

The process of making a trained model available for inference, either as a real-time API endpoint or a batch job.

A RESTful API that serves predictions from a deployed model, allowing applications to send input data and receive outputs.

The infrastructure and process of hosting a model to handle prediction requests, including scaling, monitoring, and logging.

Running predictions on a large dataset in scheduled or triggered jobs, as opposed to real-time per-request serving.

The ongoing tracking of a deployed model's performance, data drift, and operational health to ensure reliability and accuracy.

A scheduled or on-demand execution of a script or notebook in Cloudera AI, often used for data pipelines or model training.

The configuration of jobs to run at specified times or intervals (e.g., cron) or triggered by events, ensuring timely execution.

The collection and visualization of metrics, logs, and alerts to track the health and performance of sessions, jobs, and deployments.

The policies, controls, and processes that ensure data and models are used ethically, securely, and in compliance with regulations.

The mechanisms that restrict who can view, modify, or execute resources (projects, data, models) based on roles and permissions.

The secure storage and handling of sensitive information (e.g., passwords, API keys) used to access data sources and services.

The practice of assigning unique identifiers to different iterations of a model, enabling rollback and comparison.

A change in the statistical properties of input data over time that can degrade model performance, requiring monitoring and retraining.

The process of assessing a model's performance using metrics (e.g., accuracy, precision, recall) on validation or test datasets.

The creation of new input variables from raw data to improve model performance, often involving transformation and selection.

The process of cleaning, transforming, and organizing raw data into a suitable format for analysis and modeling.

The specific version of the software environment (e.g., Python 3.9, Spark 3.2) used in a session or job.