Searching CertSprout…

CertSproutGrow your skills. Collect your wins.

CertSprout is not affiliated with CompTIA, Microsoft, Cisco, ISC2, AWS or any other certification body. CertSprout reads credentials from issuers’ public records at the credential holder’s request.

For teams Privacy Terms Support About our bot Created by Samios Software Solutions

A little curiosity goes a long way

Find your next discovery.

Search certifications, learning topics, career paths, public profiles and help across CertSprout.

Search CertSprout

Browse certifications by field or career path →Provider offerings & September 2026 review notes →

Everything Certifications Mini-lessons People Career paths Study resources Exam deals App & help

Showing 20 results · Page 15 for “oci”. More results are available below.

Mini-lessons

DBX-GAI · Responsible AI evaluation

Evaluate failures for the people affected

Aggregate accuracy can hide poor performance for particular users or situations. Evaluate representative subgroups and realistic failure cases, and consider privacy, misuse and human oversig…

DBX-MLA · Databricks lakehouse concepts

Improve data quality through explicit layers

A lakehouse workflow can separate raw ingestion from validated and business-ready datasets. Each layer should have a clear contract rather than merely a different folder name. Preserve enoug…

DBX-MLA · Generative AI systems

Grounding supplies context, not a guarantee

← PreviousPage 15Next →

People search uses public names, handles and headlines. Private and unlisted profiles are excluded. Your goals are visible only to you in search.

A generative model produces output based on its learned patterns and the context it receives. Retrieval can supply relevant documents, but the model may still misunderstand or invent details…

DBX-MLA · Machine-learning engineering

Prevent leakage across the training boundary

Data leakage occurs when training or model selection uses information that would not be available at prediction time. Split data in a way that respects time and related entities, and fit lea…

DBX-MLA · Responsible AI evaluation

Evaluate failures for the people affected

Aggregate accuracy can hide poor performance for particular users or situations. Evaluate representative subgroups and realistic failure cases, and consider privacy, misuse and human oversig…

DBX-SPARK · Data pipeline design

Design pipelines for recovery and replay

A data pipeline needs defined inputs, transformations, outputs and checkpoints. Failures can leave partial results, so decide how processing resumes without losing or duplicating records. Tr…

DBX-SPARK · Data quality

Validate meaning as well as format

A value can have the right data type and still be wrong for the business. Data quality checks should cover required values, uniqueness, valid ranges and relationships, along with freshness a…

DBX-SPARK · Databricks lakehouse concepts

Improve data quality through explicit layers

A lakehouse workflow can separate raw ingestion from validated and business-ready datasets. Each layer should have a clear contract rather than merely a different folder name. Preserve enoug…

DCA · Container fundamentals

An image is a template; a container is an instance

A container image packages application files and runtime dependencies. A running container adds process state and a writable layer, while sharing the host kernel through isolation mechanisms…

DCA · Docker operations

Keep persistent data outside disposable containers

A container's writable layer is tied to that container's lifecycle. Use an appropriate volume or external service for data that must survive replacement. Configure networking and published p…

DCA · Network protection

Allow the required flow at the right boundary

… firewall tracks connection state and distinguishes associated return traffic.

DELL-ISM · Backup and recovery

A backup succeeds when the restore succeeds

Recovery point objective describes tolerable data loss measured in time; recovery time objective describes tolerable restoration time. Backup frequency influences the first, while restore sp…

DELL-ISM · Server infrastructure

Redundancy needs independent failure paths

Redundant components improve availability only when a single failure cannot disable all of them. Two power supplies connected to the same failed power source still lose power together. Trace…

DELL-ISM · Storage concepts

Choose storage by access pattern

Block storage presents volumes to an operating system, file storage organizes shared files and directories, and object storage addresses objects with metadata. The workload determines which …

DP-100 · Azure machine learning workflow

Track the experiment and the deployed behavior

A machine-learning experiment needs reproducible data, code, parameters and evaluation results. Keep training and evaluation data appropriately separate to avoid leakage. Deployment adds ano…

DP-100 · Machine-learning engineering

Prevent leakage across the training boundary

Data leakage occurs when training or model selection uses information that would not be available at prediction time. Split data in a way that respects time and related entities, and fit lea…

DP-100 · Statistics and measurement

A sample estimate includes uncertainty

…fore interpreting an average. Correlation describes association and does not, by itself, establish that one variable caused another. Worked example A sample contains 2, 4 and 9. What is its …

DP-300 · Azure SQL operations

Investigate the query before adding capacity

Slow database behavior can come from query plans, blocking, indexes, data volume or resource limits. Establish when the problem began and whether it affects one query or the whole workload. …

DP-300 · Database administration

Protect recoverability while tuning performance

Database administration balances availability, integrity, performance and recovery. Investigate workload evidence before changing indexes or resource limits, and understand the impact on wri…

DP-300 · SQL querying

Define the rows before aggregating them

SQL joins determine which rows are combined, filters determine which remain, and aggregation summarizes the result. Confirm the intended grain before calculating a total. Joining one parent …