メイン コンテンツにスキップ
Microsoft
separator
https://catalogartifact.azureedge.net/publicartifacts/ltim.ltimindtree_kenai_6_weeks_poc-c5578286-0ff7-40b4-a53b-778064d494a7/image4_LTM216X216px.png

LTM Data & AI KenAI - 6 Weeks PoC

LTIMindtree Limited

LTM KenAI redefines enterprise AI operations by unifying automation, governance, and monitoring, enabling organizations to rethink and scale machine learning with a disruptive, integrated approach.

Accelerate your MLOps journey on Microsoft Azure using KenAI through a six-week proof-of-concept engagement that operationalizes machine learning models into production while enabling model quality assurance, explainability, and monitoring.

LTM KenAI is an integrated automation framework that helps standardize and streamline the AI and machine learning lifecycle. Built on Azure-native services, it helps data science and engineering teams operationalize models faster and scale them efficiently using reusable components and automated pipelines.

KenAI enables risk and governance teams to view model quality and compliance through monitoring dashboards. Business users gain visibility into model outcomes through explanations and what-if analysis. The platform also supports collaboration across teams by providing insights into model performance, drift detection, and operational health.

Azure Components:

  • Azure DevOps
  • Azure Machine Learning
  • Azure Kubernetes Service
  • Azure Blob Storage
  • Azure Monitor
  • Azure Application Insights
  • Core Capabilities:

  • KenAI Automate: Automated pipelines for model training, inference, and drift detection using reusable lifecycle components.
  • KenAI Assure: Model quality assurance with built-in testing to measure performance and reliability.
  • KenAI Govern: Model explainability and governance to improve transparency and support responsible AI.
  • KenAI Monitor: Unified monitoring for tracking model performance, service health, and operational metrics.
  • Proof of Concept Scope:

  • Operationalization of one standard machine learning model.
  • Demonstration of model quality assurance, explainability, and monitoring.
  • Requires availability of baseline or ground truth data for drift detection.
  • Does not include model development or model-building activities.
  • 6-Week PoC Engagement Plan:

    Week 1:

  • Review current Azure Machine Learning environment and assess existing models and datasets.
  • Finalize the model and supporting artifacts for the PoC.
  • Set up metadata stores including database and schema structures.
  • Week 2 and 3:

  • Configure PoC user access, metadata schema, and required model/dataset metadata stores.
  • Migrate user base and databases to Azure SQL Database.
  • Migrate code repositories to Azure DevOps.
  • Deploy KenAI on an Azure virtual machine.
  • Define and update KPIs for model performance.
  • Week 4, 5, and 6:

  • Configure Azure Machine Learning pipelines for training, inference, and drift detection.
  • Deploy model using KenAI reusable components.
  • Perform model testing and create reliability dashboards using KenAI Assure.
  • Enable explainability using KenAI Govern.
  • Configure monitoring pipelines for drift detection and service health dashboards.
  • Outcomes:

  • Up to 2x faster time-to-market for AI and ML implementations.
  • Improved observability through business and operational monitoring.
  • Stronger adherence to responsible AI through explainability and transparency.
  • Streamlined MLOps using CI/CD, continuous training, and monitoring.
  • Client prerequisites:

  • Existing ML model selected for PoC.
  • Training/inference scripts or deployable model artifacts available.
  • Baseline dataset and, where applicable, ground-truth data available.
  • Azure subscription/resource group access available.
  • Required approvals for Azure DevOps, Key Vault, SQL/metadata store, monitoring, and storage.
  • Client SME availability for model, data, infra, and security discussions
  • Azure subscription/resource group access available.
  • 概要

    https://catalogartifact.azureedge.net/publicartifacts/ltim.ltimindtree_kenai_6_weeks_poc-c5578286-0ff7-40b4-a53b-778064d494a7/image0_1KenAIFramework.png
    https://catalogartifact.azureedge.net/publicartifacts/ltim.ltimindtree_kenai_6_weeks_poc-c5578286-0ff7-40b4-a53b-778064d494a7/image6_2BuildingBlocks.png
    https://catalogartifact.azureedge.net/publicartifacts/ltim.ltimindtree_kenai_6_weeks_poc-c5578286-0ff7-40b4-a53b-778064d494a7/image2_3SuccessStories2.png