ai powered

AI-Powered Knowledge Mining Hackathon

Students in a classroom-min

Course Description

This Hack enables participants to add intelligent search capabilities to their applications and services, leveraging artificial intelligence to extract meaningful results from data.  

This Hack simulates a real-world scenario where a travel company needs to uncover data locked up in documents and withdraw insights from that data to make key business decisions.  

During the “hacking”, participants will focus on: 

  1. Exploring ways in which Azure Search can be used as the core of a search solution.
  2. Enriching the search solution through integration with Cognitive Services, Azure Machine Learning, and custom code. 

By the end of the Hack, participants will have built out a technical solution that is a complete Azure machine learning-based intelligent search infrastructure that can interpret vast quantities of diverse data (i.e. documents, scanned images, and other digital artifacts). 

About this course


Azure Cognitive Services, Azure Functions, Question Answering, Language Understanding Service, Form Recognizer, Azure Machine Learning 


To be successful and get the most out of this Hack, it is highly recommended that participants have previous experience with either C#, JavaScript, or Python coding language and are familiar with the technologies listed above. Participants should have a basic understanding of how different services interact through APIs (Application Programing Interfaces), including REST/JSON interfaces.  

Required knowledge of Azure Fundamentals.


Challenge 1: A Question of Knowledge 

In this challenge, you will use Microsoft Question Answering cognitive service to build and publish a knowledge base.  

Learning objectives: 

  • Create a question answering knowledge base that includes suitable question and answer pairs 
  • Write code that uses the REST API to query the knowledge base 

Challenge 2: The Search Begins 

In this challenge, you will create a datasource, index, and indexer and demonstrate code that successfully retrieves the information. 

Learning objectives: 

  • Create an Azure Search Index 
  • Use the SDK or REST API to submit a range of search queries using both simple and full syntax 

Challenge 3: Expanding the Search 

In this challenge, you will update the index and demonstrate code that successfully retrieves information.  

Learning objectives: 

  • Add built-in cognitive skills to an Azure Search index to return: 
  • Key Phrases 
  • Entities (including links) 
  • Sentiment (especially reviews) 

Challenge 4: Getting the Full Picture 

In this challenge, you will use the OCR skill to expand the index to extract AI-generated descriptions of images embedded in documents.  

Learning objectives: 

  • Add built-in cognitive skills to an Azure Search index to return: 
  • Image Descriptions and Tags 
  • OCR extracted Text 

Challenge 5: What is the Frequency? 

In this challenge, you will create a web API custom skill for your Azure Search index. 

Learning objectives: 

  • Create a custom skill to find the top ten most frequent words 
  • Incorporate your custom skill into your web content index 

Challenge 6: The Search for Relevance 

In this challenge, you will create return search results based on synonyms and display suggestions and autocomplete options.  

Learning objectives: 

  • Add Synonyms to Azure Search index to ensure relevant results 
  • Implement query suggestions / autocomplete 
  • Add scoring profiles that boost documents in results 

Challenge 7: Knowledge preservation 

In this challenge, you will modify the index skillset to generate knowledge store assets. Browse the blobs and tables in the knowledge store using the Storage Explorer interface.  

Learning objectives: 

  • Implement knowledge stores. 

Challenge 8: Finding Your Form 

In this challenge, you will train a Form Recognizer model and integrate into a new Azure Search index.  

Learning objectives: 

  • Create a custom skill that calls the Forms Understanding Service 

Challenge 9: Use Your Intelligence 

In this challenge, you will publish a machine learning model as a web service to predict claim probability.  

Learning objectives: 

  • Build a custom skill that consumes a custom machine learning model 

Value Proposition

  • This Hack provides a real-world context of Azure services for knowledge mining with Azure Search. 
  • Learn common knowledge mining and intelligent search scenarios. 
  • AI-Powered Knowledge Mining helps businesses make better decisions through more robust data extraction and analysis. 
  • Improved search functionality to decrease time-to-find relevant data. 

Technical Scenarios

  • Search – Improving the ability to ingest data from various sources, intelligently sift through that data using core Azure AI services like Azure Search and Cog Services, and then analyze that data 
  • Data modeling – Create custom ML models that learn from data and continuously improve search 


  • Target Audience:  
  • Microsoft – CSE, CSA, GBB, ATT, SE, TPM
  • Customer – App Developers, ML/AI Engineers
  • Target verticals: Retail, Cross-Industry
  • Customer profile:
  • Customers looking to extract insights from vast quantities of diverse data 
  • Customers looking for intuitive search capabilities to help them find and make sense of relevant information that they need to make key business decisions. 


  • 3 days

Ready to get started?