# MLRun.org --- ## Pages - [Homepage new](https://www.mlrun.org/): MLRun offers an integrative approach to manage your ML pipelines from early development through management in your production environment. - [Blog](https://www.mlrun.org/blog/): Stay updated with MLRun’s latest tutorials, releases, and insights on MLOps and Gen AI orchestration to enhance your ML workflows. - [Choose a deployment option:](https://www.mlrun.org/choose-a-deployment-option/): Explore MLRun deployment options: run on your laptop, on Kubernetes, on AWS, or choose our managed enterprise service. --- ## Posts - [Fine-Tuning in MLRun: How to Get Started](https://www.mlrun.org/blog/fine-tuning-in-mlrun-how-to-get-started/): How MLRun simplifies and accelerates fine-tuning workflows. See two practical, hands-on examples, which you can easily follow and replicate. - [MLRun v1.8 Release: with Smarter Model Monitoring, Alerts and Tracking](https://www.mlrun.org/blog/mlrun-v1-8-release-with-smarter-model-monitoring-alerts-and-tracking/): MLRun v1.8 adds features to make LLM and ML evaluation and monitoring more accessible, practical and resource-efficient. - [Bringing (Gen) AI from Laptop to Production with MLRun](https://www.mlrun.org/blog/bringing-gen-ai-from-laptop-to-production-with-mlrun-2/): A gen AI copilot is an interactive AI assistant that is designed to amplify human capabilities. Here's how to build one using MLRun. - [MLRun Customer Support Gen AI Copilot](https://www.mlrun.org/blog/mlrun-customer-support-gen-ai-copilot/): A gen AI copilot is an interactive AI assistant that is designed to amplify human capabilities. Here's how to build one using MLRun. - [How to Connect MLRun to an External Monitoring Application](https://www.mlrun.org/blog/how-to-connect-mlrun-to-an-external-monitoring-application/): Integrating MLRun with an external monitoring application is simple and straightforward. Here’s how it works. - [Launching MLRun 1.7: Gen AI and LLM Monitoring](https://www.mlrun.org/blog/launching-mlrun-1-7-gen-ai-and-llm-monitoring/): Introducing MLRun 1.7: enhanced Gen AI deployment with flexible monitoring, unstructured data tracking, and customizable endpoint metrics. - [LLM as a Judge: Practical Example with Open-Source MLRun](https://www.mlrun.org/blog/llm-as-a-judge-practical-example-with-open-source-mlrun/): Learn some practical example of operationalizing and de-risking an LLM as a Judge in with the open-source MLRun platform. - [How to Operationalize Your Own Customized Application for Monitoring LLMs with MLRun](https://www.mlrun.org/blog/operationalize-your-own-customized-application-for-monitoring-llms/): Learn how to build and operationalize a customized monitoring application for LLMs using MLRun to boost performance and insights. - [Open Source MLOps and LLMOps Orchestration with MLRun: Quick Start Tutorial](https://www.mlrun.org/blog/open-source-mlops-and-llmops-orchestration/): Quickly start with MLRun: create projects, prepare data, train models, and deploy with this comprehensive tutorial. - [Deploying Hugging Face LLM Models with MLRun](https://www.mlrun.org/blog/deploying-hugging-face-llm-models-with-mlrun/): Deploy Hugging Face LLMs with MLRun: automate data preparation, model tuning, validation, and scalable deployment. - [Tutorial: Build a Smart Call Center Analysis Gen AI App with MLRun, Gradio and SQLAlchemy](https://www.mlrun.org/blog/build-a-smart-call-center-analysis-gen-ai-app/): Build a smart call center analysis Gen AI app using MLRun, Gradio, and SQLAlchemy with this step-by-step tutorial. --- ## Resources - [Bring Data Science to Life: Develop on SageMaker | Deploy with MLRun on Iguazio (DEMO)](https://www.mlrun.org/blog/resources/aws-and-iguazio-bring-data-science-to-life/): AWS and Iguazio Bring Data Science to Life: Develop on SageMaker --- ## News - [How to mask PII before LLM training with MLRun](https://www.mlrun.org/blog/news/how-to-mask-pii-before-llm-training-with-mlrun/): The PII Recognizer is an open source function that can detect PII data in datasets and anonymize the PII entity... - [Demo: Smart Call Center Analysis App](https://www.mlrun.org/blog/news/demo-smart-call-center-analysis-app/): In this demo we will be showcasing how we used LLMs to turn call center conversation audio files of customers... - [MLOps for Generative AI with MLRun](https://www.mlrun.org/blog/news/mlops-for-generative-ai-with-mlrun/): The influx of new tools like ChatGPT spark the imagination and highlight the importance of Generative AI and foundation models... - [How to train on Databricks and deploy with MLRun](https://www.mlrun.org/blog/news/how-to-train-on-databricks-and-deploy-with-mlrun/): Here’s how to use Databricks for model training and tracking while using MLRun to deploy model serving. Read the full... - [Kubeflow Vs. MLflow Vs. MLRun: Which One is Right for You?](https://www.mlrun.org/blog/news/kubeflow-vs-mlflow-vs-mlrun-which-one-is-right-for-you/): The open source ML tooling ecosystem has become vast in the last few years, with many tools covering different aspects... - [Deploying Your Hugging Face Models to Production at Scale with MLRun](https://www.mlrun.org/blog/news/deploying-your-hugging-face-models-to-production-at-scale-with-mlrun/): Using Hugging Face and MLRun together significantly shortens the model development, training, testing, deployment and monitoring process. By getting your... - [Using MLRun to Automate Logging & Tracking for ML](https://www.mlrun.org/blog/news/using-mlrun-to-automate-logging-tracking-for-ml/): From AutoML to AutoMLOps: This post shows you how to automate engineering tasks like logging and tracking, so that your... - [How to Deploy an MLRun Project in a CI/CD Process with Jenkins Pipeline](https://www.mlrun.org/blog/news/how-to-deploy-an-mlrun-project-in-a-ci-cd-process-with-jenkins-pipeline/): This article walks you through the steps to run a Jenkins server in docker and deploy the MLRun project using... - [How to Build an AI App in Under 20 Minutes](https://www.mlrun.org/blog/news/how-to-build-an-ai-app-in-under-20-minutes/): Here’s how to build simple AI applications that leverage MLRun and some pre-built ML models and allow you to interact... - [Machine Learning Experiment Tracking: From Zero to Hero in 2 Lines of Code](https://www.mlrun.org/blog/news/machine-learning-experiment-tracking-from-zero-to-hero-in-2-lines-of-code/): Here’s a simple tutorial on how to turn your existing model training code into an MLRun job and get the... - [MLRun v1.0.0](https://www.mlrun.org/blog/news/mlrun-v1-0-0/): MLRun v0. 10 is now generally available. This release includes: Snowflake is now supported as a datasource for the feature... - [How to build an AI app with MLRun in 5 min](https://www.mlrun.org/blog/news/how-to-build-an-ai-app-with-mlrun-in-5-min/): A simple tutorial showing how to build an image classifier AI app in 5 minutes with MLRun, including a look... - [MLRun 0.10](https://www.mlrun.org/blog/news/mlrun-0-10/): MLRun v0. 10 is now generally available. This release includes: New project-level menu. When accessing a project in the Projects page,... - [MLRun @ ODSC: Git-based CI/CD for ML](https://www.mlrun.org/blog/news/mlrun-odsc-git-based-ci-cd-for-ml/): In this session, Yaron Haviv, MLRun Co-Founder, discusses how to enable continuous delivery of machine learning to production using Git-based... - [MLRun 0.9 Release](https://www.mlrun.org/blog/news/mlrun-0-9-release/): MLRun v0. 9 is now generally available. This release includes: MySQL is now the underlying database in MLRun, replacing SQLite. Feature... - [MLRun 0.8 Release](https://www.mlrun.org/blog/news/mlrun-0-8-release/): MLRun v0. 8 is now generally available. This release includes: Projects now have members (users, user groups), thereby controlling access... - [MLRun 0.7.0](https://www.mlrun.org/blog/news/mlrun-0-7-0/): MLRun version 0. 7. 0 and 0. 7. 1 are now generally available. These releases includes the following: Automate the... - [MLRun 0.6.5 release](https://www.mlrun.org/blog/news/mlrun-0-6-5-release/): MLRun version 0. 6. 5 is now generally available. The release includes the following: Supports node selectors when running a... - [#MLOpsforGood Hackathon Projects](https://www.mlrun.org/blog/news/mlopsforgood-hackathon-projects/): Take a look at a few of the inspiring projects from our first-ever hackathon: ICU-OPS: Help frontline clinicians triage patients... - [MLRun 0.6.4 Release](https://www.mlrun.org/blog/news/mlrun-0-6-4-release/): Check out the latest release of MLRun, with several new features and enhancements. Feature store updates: Parquet partitioning is now... - [MLRun 0.6.3 Release](https://www.mlrun.org/blog/news/mlrun-0-6-3-release/): Check out the latest release of MLRun, with updates to the feature store, and more functionality from the UI. Feature... --- # # Detailed Content ## Pages --- ## Posts > Build a smart call center analysis Gen AI app using MLRun, Gradio, and SQLAlchemy with this step-by-step tutorial. - Published: 2024-06-13 - Modified: 2025-02-27 - URL: https://www.mlrun.org/blog/build-a-smart-call-center-analysis-gen-ai-app/ - Categories: Uncategorized Analysis - Generating a table with the call summary, its main topic, customer tone, upselling attempts and more: --- --- ## Resources - Published: 2021-01-31 - Modified: 2021-04-05 - URL: https://www.mlrun.org/blog/resources/aws-and-iguazio-bring-data-science-to-life/ AWS and Iguazio Bring Data Science to Life: Develop on SageMaker --- --- ## News - Published: 2024-01-21 - Modified: 2024-01-21 - URL: https://www.mlrun.org/blog/news/how-to-mask-pii-before-llm-training-with-mlrun/ The PII Recognizer is an open source function that can detect PII data in datasets and anonymize the PII entity in the text. Here's how it works: https://www. iguazio. com/blog/how-to-mask-pii-before-llm-training/ --- - Published: 2023-09-21 - Modified: 2024-01-21 - URL: https://www.mlrun.org/blog/news/demo-smart-call-center-analysis-app/ In this demo we will be showcasing how we used LLMs to turn call center conversation audio files of customers and agents into valueable data in a single workflow orchastrated by MLRun. MLRun will be automating the entire workflow, auto-scale resources as needed and automatically log and parse values between the workflow different steps. By the end of this demo you will see the potential power of LLMs for feature extraction, and how easy it is being done using MLRun! https://www. youtube. com/watch? v=Bx_swL4GeiI --- - Published: 2023-06-01 - Modified: 2023-06-01 - URL: https://www.mlrun.org/blog/news/mlops-for-generative-ai-with-mlrun/ The influx of new tools like ChatGPT spark the imagination and highlight the importance of Generative AI and foundation models as the basis for modern AI applications. However, the rise of generative AI also brings a new set of MLOps challenges. Challenges like handling massive amounts of data, large scale computation and memory, complex pipelines, transfer learning, extensive testing, monitoring, and so on. In this 9 minute demo video, we share MLOps orchestration best practices and explore open source technologies available to help tackle these challenges. We show ways to enable your team to automate the continuous integration and deployment (CI/CD) of foundation models and transformers, along with the application logic, in production, and how to use GPUs to maximize application performance while protecting your investment in AI infrastructure. We share tips on what to look out for and how to make the whole process efficient, effective and collaborative. --- - Published: 2023-05-15 - Modified: 2023-05-15 - URL: https://www.mlrun.org/blog/news/how-to-train-on-databricks-and-deploy-with-mlrun/ Here's how to use Databricks for model training and tracking while using MLRun to deploy model serving. Read the full tutorial here. --- - Published: 2023-03-28 - Modified: 2023-05-15 - URL: https://www.mlrun.org/blog/news/kubeflow-vs-mlflow-vs-mlrun-which-one-is-right-for-you/ The open source ML tooling ecosystem has become vast in the last few years, with many tools covering different aspects of the complex and expansive process of building, deploying and managing AI in production. Some tools overlap in their capabilities while others complement each other nicely. In part because AI/ML is still an emerging and ever-evolving practice, the messaging around what all these tools can accomplish can be quite vague. In this article, we’ll dive into three tools to better understand their capabilities, the differences between them, and how they fit into the ML lifecycle. Read the blog here. --- - Published: 2022-11-15 - Modified: 2023-05-15 - URL: https://www.mlrun.org/blog/news/deploying-your-hugging-face-models-to-production-at-scale-with-mlrun/ Using Hugging Face and MLRun together significantly shortens the model development, training, testing, deployment and monitoring process. By getting your models to production faster, you can answer business needs faster while saving resources. Read the blog here, and watch the full demo here. --- - Published: 2022-10-03 - Modified: 2022-10-03 - URL: https://www.mlrun.org/blog/news/using-mlrun-to-automate-logging-tracking-for-ml/ From AutoML to AutoMLOps: This post shows you how to automate engineering tasks like logging and tracking, so that your code is automatically ready for production. Read more here. --- - Published: 2022-08-16 - Modified: 2022-08-16 - URL: https://www.mlrun.org/blog/news/how-to-deploy-an-mlrun-project-in-a-ci-cd-process-with-jenkins-pipeline/ This article walks you through the steps to run a Jenkins server in docker and deploy the MLRun project using Jenkins pipeline. --- - Published: 2022-08-16 - Modified: 2022-08-16 - URL: https://www.mlrun.org/blog/news/how-to-build-an-ai-app-in-under-20-minutes/ Here's how to build simple AI applications that leverage MLRun and some pre-built ML models and allow you to interact with a UI in Streamlit to visualize the results. Check out the full tutorial here and then build your own! --- - Published: 2022-06-30 - Modified: 2022-06-30 - URL: https://www.mlrun.org/blog/news/machine-learning-experiment-tracking-from-zero-to-hero-in-2-lines-of-code/ Here's a simple tutorial on how to turn your existing model training code into an MLRun job and get the benefit of all the experiment tracking, plus more. --- - Published: 2022-05-30 - Modified: 2022-05-30 - URL: https://www.mlrun.org/blog/news/mlrun-v1-0-0/ MLRun v0. 10 is now generally available. This release includes: Snowflake is now supported as a datasource for the feature storeConsumer group streaming report has been addedAn option to configure pods priority and spot instances from the UI has been addedMasking of sensitive data in the UI has been enhancedConfiguration of CPU, GPU, and memory default limits for user jobs is now supported --- - Published: 2022-05-29 - Modified: 2022-05-29 - URL: https://www.mlrun.org/blog/news/how-to-build-an-ai-app-with-mlrun-in-5-min/ A simple tutorial showing how to build an image classifier AI app in 5 minutes with MLRun, including a look under the hood to see how everything is connected. --- - Published: 2022-03-09 - Modified: 2022-03-21 - URL: https://www.mlrun.org/blog/news/mlrun-0-10/ MLRun v0. 10 is now generally available. This release includes: New project-level menu. When accessing a project in the Projects page, the new menu on the left provides access to all aspects of the project, including the Project Settings. Improved UI performance when fetching objects from the MLRun DB. This relates mainly to the jobs and projects pages. The MLRun client and images are compatible with minor MLRun releases that are released in the subsequent 6 months. The internal MLRun database was upgraded to MySQL 8. 0. Spark is now supported as an engine for fetching data from a feature vector (Tech preview). Project-level k8s secrets can now be managed by API and SDK. Supports service account for MLRun functions. --- - Published: 2022-02-28 - Modified: 2022-05-29 - URL: https://www.mlrun.org/blog/news/mlrun-odsc-git-based-ci-cd-for-ml/ In this session, Yaron Haviv, MLRun Co-Founder, discusses how to enable continuous delivery of machine learning to production using Git-based ML pipelines (Github Actions) with hosted training and model serving environments. He touched upon how to leverage Git to solve rigorous MLOps needs: automating workflows, reviewing models, storing versioned models as artifacts, and running CI/CD for ML. He also covered how to enable controlled collaboration across ML teams using Git review processes and how to implement an MLOps solution with MLRun and hosted ML platforms. The session includes a live demo. --- - Published: 2022-01-04 - Modified: 2022-01-24 - URL: https://www.mlrun.org/blog/news/mlrun-0-9-release/ MLRun v0. 9 is now generally available. This release includes: MySQL is now the underlying database in MLRun, replacing SQLite. Feature store updates: Spark is now certified as an engineData can be ingested incrementallyGoogle BigQuery is now supported as a data source UI: The new Kubeflow Pipeline report is included in the Projects pageProjects have a new Settings page that includes secrets management --- - Published: 2021-11-02 - Modified: 2021-12-08 - URL: https://www.mlrun.org/blog/news/mlrun-0-8-release/ MLRun v0. 8 is now generally available. This release includes: Projects now have members (users, user groups), thereby controlling access to view and manage the project. (Available with the Iguazio MLOps Platform 3. 2) The new Settings dialog in the Projects page has two tabs: General, and Secrets, to manage general settings and project secrets. The project secrets are automatically added to MLRun jobs and nuclio functions. A feature set can now be created in the Project page of the UI. When creating a new function in the Projects page, you can now select the type of volume mount: auto, manual, none. Access Keys are now specific to the data plane, the control plane, or both (Previously there was one type of access key). Users can create a feature store that is based on Kafka as a source. --- - Published: 2021-09-09 - Modified: 2021-10-21 - URL: https://www.mlrun.org/blog/news/mlrun-0-7-0/ MLRun version 0. 7. 0 and 0. 7. 1 are now generally available. These releases includes the following: Automate the model monitoring deploymentSupport ingestion of deltas for a scheduled feature setAdd a new wizard for creating an MLRun functionUsability enhancements around working with projects, function , volume mounts, pipeline etc. . Fixed window returns the last closed window --- - Published: 2021-08-08 - Modified: 2021-08-16 - URL: https://www.mlrun.org/blog/news/mlrun-0-6-5-release/ MLRun version 0. 6. 5 is now generally available. The release includes the following:  Supports node selectors when running a new jobFeature Store updates: Improved ingestion and aggregation logic and performance Model UI now includes information about the feature vectorKubernetes secrets provider is now supportedEnables mpijob as an MLRun runtimeAdditional logging capabilities of PyTorch and Tensorflow models using TensorBoardA new option to create new functions via the UI --- - Published: 2021-07-19 - Modified: 2021-08-16 - URL: https://www.mlrun.org/blog/news/mlopsforgood-hackathon-projects/ Take a look at a few of the inspiring projects from our first-ever hackathon: ICU-OPS: Help frontline clinicians triage patients in ICUs by rapidly assessing a patient's overall health for informed clinical decisions that will improve patient outcomes and relieve COVID19 ICU overload. Deepfake Shield: Fighting deepfakes with deep-learning. COVID-19 Detection through Radiography images of Lung CT: COVID-19 detection using Radiography images based on ResNet50V2 architecture. --- - Published: 2021-06-06 - Modified: 2021-10-13 - URL: https://www.mlrun.org/blog/news/mlrun-0-6-4-release/ Check out the latest release of MLRun, with several new features and enhancements. Feature store updates:Parquet partitioning is now supportedRead offline data with time rangeData ingestion using Spark Reduce the docker images size for faster delivery over the networkA new CLI option to submit functions/projects to the clusterDeploy model via the UISupport code archives Read the full release notes here. --- - Published: 2021-05-13 - Modified: 2021-10-13 - URL: https://www.mlrun.org/blog/news/mlrun-0-6-3-release/ Check out the latest release of MLRun, with updates to the feature store, and more functionality from the UI. Feature store updated:Support for multiple ingestion engines: Pandas, Spark, StoreyIngestion now supports filtering by dateFeature sets now include code usage examples Model monitoring has been improved and now includes feature analysis and drift analysis informationA graph execution can be split to different paths. Run a new job via the UI based on a previous job and change some of the parameters. Jobs can now be aborted via the UI. View the pods of running jobs as well as the details of each pod. Job filtering now includes a filter by status and by the start time. A feature vector can now be created via the UI See the full release notes here --- ---