Free Udemy Course: DP-100: Designing and Implementing a Data Science Solution

Skills at a glanceDesign and prepare a machine learning solution (20–25%)Explore data, and run experiments (20–25%)Train and deploy models (25–30%)Optimize language models for AI applications (25–30%)Design and prepare a machine learning solution (20–25%)Design a machine learning solutionIdentify the structure and format for datasetsDetermine the compute specifications for machine learning workloadSelect the development approach to train a modelCreate and manage resources in an Azure Machine Learning workspaceCreate and manage a workspaceCreate and manage datastoresCreate and manage compute targetsSet up Git integration for source controlCreate and manage assets in an Azure Machine Learning workspaceCreate and manage data assetsCreate and manage environmentsShare assets across workspaces by using registriesExplore data, and run experiments (20–25%)Use automated machine learning to explore optimal modelsUse automated machine learning for tabular dataUse automated machine learning for computer visionUse automated machine learning for natural language processingSelect and understand training options, including preprocessing and algorithmsEvaluate an automated machine learning run, including responsible AI guidelinesUse notebooks for custom model trainingUse the terminal to configure a compute instanceAccess and wrangle data in notebooksWrangle data interactively with attached Synapse Spark pools and serverless Spark computeRetrieve features from a feature store to train a modelTrack model training by using MLflowEvaluate a model, including responsible AI guidelinesAutomate hyperparameter tuningSelect a sampling methodDefine the search spaceDefine the primary metricDefine early termination optionsTrain and deploy models (25–30%)Run model training scriptsConsume data in a jobConfigure compute for a job runConfigure an environment for a job runTrack model training with MLflow in a job runDefine parameters for a jobRun a script as a jobUse logs to troubleshoot job run errorsImplement training pipelinesCreate custom componentsCreate a pipelinePass data between steps in a pipelineRun and schedule a pipelineMonitor and troubleshoot pipeline runsManage modelsDefine the signature in the MLmodel filePackage a feature retrieval specification with the model artifactRegister an MLflow modelAssess a model by using responsible AI principlesDeploy a modelConfigure settings for online deploymentDeploy a model to an online endpointTest an online deployed serviceConfigure compute for a batch deploymentDeploy a model to a batch endpointInvoke the batch endpoint to start a batch scoring jobOptimize language models for AI applications (25–30%)Prepare for model optimizationSelect and deploy a language model from the model catalogCompare language models using benchmarksTest a deployed language model in the playgroundSelect an optimization approachOptimize through prompt engineering and prompt flowTest prompts with manual evaluationDefine and track prompt variantsCreate prompt templatesDefine chaining logic with the prompt flow SDKUse tracing to evaluate your flowOptimize through Retrieval Augmented Generation (RAG)Prepare data for RAG, including cleaning, chunking, and embeddingConfigure a vector storeConfigure an Azure AI Search-based index storeEvaluate your RAG solutionOptimize through fine-tuningPrepare data for fine-tuningSelect an appropriate base modelRun a fine-tuning jobEvaluate your fine-tuned model