Announcing the winners of the Microsoft Fabric and AI Learning Hackathon!

We’re excited to announce the winners of the Microsoft Fabric and AI Learning Hackathon!

The Microsoft Fabric and AI Learning Hackathon:

In this Microsoft Fabric Focused Hackathon event, we partnered with DevPost to challenge the world to build the next wave of innovative AI powered data analytics applications with Microsoft Fabric! The event was announced just ahead of the European Microsoft Fabric Community Conference on September 16th and continued through to November 12th, just ahead of Microsoft Ignite. We had some advance notice of the recent Fabric aligned announcements that came from these events, and you’ll notice that our prize categories aligned to this in a very thoughtful way. We’ve opened the doors even wider for database and developer use cases in the product!

To ensure that hackers had access to the latest and greatest AI services from Microsoft we provided an Azure Open AI proxy service that allowed hackers to incorporate Azure Open AI without the need for an active Azure subscription. Over the course of the Hackathon, hackers made over 88,000 requests to the Azure Open AI proxy!

To further support readiness for all participants, we produced 10 livestreams designed to train and prepare our entrants for the competition. We also hosted office hours on the DevPost Hackathon Discord and AMA events to ensure hackers had regular touchpoints with judges while working on their project. By the conclusion of the event​, we had onboarded 4,550+ participants to compete for prizes across six categories. They are listed along with links to the relevant livestream accompanying each category:

• Grand prize – Best of the best! This category rewards the solution that meets all judging criteria, wows the judges, has potential real-world value, and demonstrates innovative use of Artificial Intelligence in Microsoft Fabric.
• Best Microsoft Fabric + AI Innovation – A solution that solves a real-world problem with Microsoft Fabric by implementing any concept or collection of concepts related to Artificial Intelligence.
• Best use of Real-Time Intelligence in Microsoft Fabric – A solution that uses Real-Time Intelligence (RTI) in Microsoft Fabric to solve a real-world problem.
• Best Azure Database for PostgreSQL Integration – A real-world solution that uses Azure Database for PostgreSQL as an OLTP main transactional database and the transactional data has to be offloaded to Fabric for near real-time analytics.
• Best SQL And AI Integration​ – A real-world solution that built on Microsoft Fabric that leverages Azure SQL Database or SQL Server and augmented by AI capabilities.
• Best Azure Cosmos DB + Microsoft Fabric Integration – A real-world solution that is built on Microsoft Fabric, leveraging Azure Cosmos DB.

The winners

We received 83 Hackathon project submissions from hackers all across the world! Judging was as difficult as ever, with a total of only 6 lucky teams able to secure the title of winner in their respective category.

The judging process considered completion of an online learning and skilling challenge, overall alignment to the corresponding Hackathon category, degree of innovation/impact to the real-world, clarity of documentation and reproducibility, and overall quality of a hacker-provided video demonstration for all submissions.

We were very excited to see so many high-quality submissions, some of which preluded feature announcements at Microsoft Ignite, tackling real-world problems in innovative and exciting ways. While we can only choose one winning team per category, we will also be highlighting standout honorable mentions for each category. Without further ado, here the winners of the Microsoft Fabric and AI Learning Hackathon!

🏆 Grand prize: Data-Driven Intelligence with Microsoft Fabric and OpenAI

Our inspiration came from the ongoing challenge that companies face in making sense of data scattered across various SaaS platforms. Sales directors often struggle to extract clear insights from an overload of charts across sales, marketing, and customer relationship data—a process that can take days or even weeks to interpret. We saw an opportunity to streamline data management and empower businesses with faster, more informed decision-making through the capabilities of Microsoft Fabric and Azure OpenAI.

Our project centralises data from multiple sources (HubSpot CRM, Azure SQL Database, AWS S3) into a unified platform using Microsoft Fabric. Leveraging Azure OpenAI’s large language models (LLMs), it generates actionable insights, such as sales trends, performance drivers, and strategic recommendations, all presented through an intuitive PowerBI report.

Project Repository URL

https://github.com/Srujan1993/datadabblers

Project Video

🏅 Best Microsoft Fabric + AI Innovation: InfraGen

The need for precise, high-quality datasets is critical in AI development. We saw an opportunity to leverage Microsoft Fabric’s powerful ecosystem to create a solution that automates and customizes real world multimodal dataset generation, helping developers save time and increase the quality of their AI models.

InfraGen is a cloud-native data pipeline that automatically generates diverse, user-specific image datasets. It combines structured and unstructured data, AI-enhanced labeling, and multimodal integration to produce high-quality datasets ready for model training, testing, and validation.

Project Repository URL: https://github.com/58191554/InFreGen

Project Video

🏅 Best use of Real-Time Intelligence in Microsoft Fabric: Real-Time Intelligence in Microsoft Fabric: MeteoWatch

We brainstormed around use cases solvable with open data that we could tackle using real-time intelligence technologies, as none of us had experience with it and were eager to try it using Microsoft Fabric. We all work at Deutsche Bahn, but unfortunately, there isn’t much open railway data available. However, there is actually a lot of open aviation data! So, we began with the question, ‘What can you do with OpenSky Network data?’ Then we discovered open data on weather hazards through OpenAviation, as well as route and vehicle reference data on ADSDB. We reached out to ADSDB and were granted permission to use their data for the hackathon. Then we put it all together: Why not build a monitoring and alerting system based on Fabric Eventhouse and use Azure OpenAI to customize alert messages?

The system consumes data from OpenSky Network, ADSDB, and AviationWeather. Using this data, MeteoWatch knows the current position of aircraft and their flight routes. The system is also aware of areas where weather hazards, known as SIGMETs (Significant Meteorological Information), are reported. MeteoWatch calculates whether a flight will be affected by a SIGMET and creates a natural language warning message, which could be tailored and sent to a specific aircraft. Additionally, MeteoWatch includes a dashboard where stakeholders such as air control or pilots can view the flight route, relevant weather hazards, and other useful information.

Project Repository URL: https://github.com/MeteoWatch/MeteoWatch/tree/main

Project Video

🏅 Best Azure Database for PostgreSQL Integration: Real-Time Equipment Monitoring and Predictive Maintenance

The inspiration for this project stemmed from the need to continuously monitor the health of industrial equipment using real-time sensor data. Sensors on devices capture metrics like temperature and vibration levels every minute, and by analyzing this data, we can predict failures before they occur. This can drastically reduce downtime and improve operational efficiency. Leveraging Azure’s cloud ecosystem and modern data architecture, we aimed to build a system that enables proactive maintenance and equips decision-makers with the data they need to ensure smooth operations.

This project focuses on real-time monitoring of equipment health through sensor data. The core functionality includes:

• Capturing real-time temperature and vibration data from devices via Azure Database for PostgreSQL with Change Data Capture (CDC) enabled.
• Streaming this data to a KQL database for real-time analysis and storage.
• Using a snapshot table to maintain the latest device status using Medallion Architecture for efficient reporting.
• Providing an interactive Power BI report that visualizes the device health, status trends, and sensor readings, enabling stakeholders to track equipment performance in real-time and take action proactively.

Project Repository URL: https://github.com/revathi2592/Real-Time-Equipment-Monitoring-and-Predictive-Maintenance

Project Video

🏅 Best SQL And AI Integration: Hiking Alerts

When viewing and planning tracks, popular hiking apps do not display current information by official institutions for hiking tracks in a standard format. Instead, those apps heavily rely on their respective communities updating the information manually.

This relevant information for hikers from institutions might include dangers, weather impacts, closures etc (in the following called events). When planning hiking tours solely via apps, there is a risk of overseeing important events.

Microsoft’s AI and Fabric services enable to standardize the events issued by different institutions in a global context and connect the events with geospatial data. The purpose of the website and the API is to provide transparency on the accessibility of hiking tracks, making hiking safer and a more enjoyable experience (https://www.hiking-alerts.org).

Project Repository URL: https://github.com/lukas8920/hikealerts

Project Video

🏅 Best Azure Cosmos DB + Microsoft Fabric Integration: Research of Research

The vast amount of academic research available today is both a treasure trove of knowledge and a challenge to navigate. We wanted to create a tool that empowers researchers by streamlining the discovery process, automating data handling, and providing relevant insights at the right time. The goal was to make academic exploration as efficient and insightful as possible, allowing researchers to focus on their work without getting bogged down in tedious data searches.

Research of Research is an automated platform that manages research queries, gathers and organizes relevant research data, and delivers actionable insights directly to users. Leveraging multiple APIs and AI-driven tagging, the system fetches, processes, and enriches metadata from scholarly databases. It updates reports dynamically in Power BI and sends them to users, ensuring they have access to the latest academic insights. Additionally, the platform continuously updates its AI-powered search index, making it a powerful tool for discovering new information with ease.

Project Repository URL: https://github.com/choisiulun1/Research_Of_Research

Project Video

💝Honorable Mentions

We received many high-quality submissions and encourage you to take a look at all of the hackathon entries to see if there is something relevant to your line of work or interests. Each of these submissions has provided value not just for the hackers but for anyone interested in looking into how their particular problems were solved using Microsoft Fabric workloads along with AI methodologies. There’s more for you to explore, check out our honorable mentions!

🪙Book Recommendations AI Assistant

The Microsoft Fabric Platform made a huge impact as a unified SaaS offering with a well-established set of integration capabilities. The inspiration came from the idea to create an AI powered recommendations engine Web App, where developers can work from a single platform for all the aspects of the build lifecycle.

This recommendations AI Assistant hosted in Azure Web Apps. The Data is one hand user SignUp details with Book Genre preferences stored in Azure SQL, and a Books Dataset stored in Fabric OneLake. Two Separate pipelines create and update on schedule Vector Indexes within Azure AI Search, where embeddings are generated from Azure OpenAI, using Jupyter Notebooks. The Notebooks perform also data cleaning and transformation, as well as incorporate API rate Limiting backoff procedures for successful embeddings generation. Users are able to login to the Web APP with username\password where a React Frontend allows them to interact with the Assistant, a NodeJS backend, that recognizes keywords like recommendation or rating and provides answers to the users. When new users sign up a Reflex is activated by CDC on Azure SQL, and it fires up a new indexing process for only the new users. The User is updated in near real time with a polling mechanism and a new field created in Azure SQL upon embeddings and index update.

Project Repository URL: https://github.com/passadis/ai-assistant

Project Video

🪙LinkedIn Job Market Analysis in Sweden

This project is designed to assist individuals who are struggling to find jobs in Sweden by providing insights into the job market. By analyzing job postings, it reveals which tools, skills, and requirements are most in demand by employers, which is particularly beneficial for those who may not have a strong understanding of the market, such as new graduates or career changers. Additionally, the project offers city-level insights, which can be especially useful for newcomers or those considering relocating to Sweden. It provides a clearer picture of which cities have the most job opportunities, making it easier for job seekers to align their job search with current market trends.

This project is an in-depth analysis designed to derive insights from LinkedIn job posts within Sweden on an hourly basis. It utilizes a robust architecture integrating various technologies and services to capture, store, process, and analyze jobs posted on LinkedIn on an hourly basis. Additionally, it incorporates population data at the city level to provide a comprehensive view of the Swedish job market dynamics.

Project Repository URL: https://github.com/AnasMofleh/Linkedin_jobs_datalake

Project Video

🪙CDC LLM Data Architecture with OpenSource and Fabric
As billions of data get generated daily having a robust system that can manage such an influx of data is of high importance. With OpenSource technology achieving such a process is possible but the question arises, how can this be done also on an enterprise level for business users?

Having a solution that can stream data in real-time to OpenSource technology like Kafka and also doing the same in Microsoft Fabric to create an LLM solution that can help businesses make quick and fast decisions.

A real-time data stream collects data from a producer and stores it in an Azure PostgreSQL Database. An LLM is built out of the stored data to respond to user queries (RAG Solution).

Project Repository URL: https://github.com/kiddojazz/CDC_Stream_Kafka_Fabric

Project Video

🪙TrendTrackr

TrendTrackr was inspired by a passion for open-source and the desire to uncover hidden stories behind GitHub activity. GitHub is a vibrant community where millions of developers come together to create, share, and innovate. We wanted to build a solution that could help us understand these dynamics: which repositories are trending, which languages are growing, and who are the top contributors driving innovation. With the power of Microsoft Fabric and Azure services, we set out to build a comprehensive analytics tool to provide insights into the evolving ecosystem of GitHub.

This journey taught us many lessons in data engineering, cloud services, and the power of real-time analytics:

• We mastered the integration of GitHub Actions for automated, real-time data collection, exploring how to seamlessly fetch GitHub events.
• Leveraging Azure Data Factory and Dataflow Gen2 for efficient ETL (Extract, Transform, Load) operations helped us understand how to bring agility and reliability to our data pipeline.
• We explored Microsoft Fabric Notebooks for deeper data analysis, crafting SQL queries that converted raw event data into actionable insights.
• Finally, Power BI became our visual storytelling tool, where we learned how to make data come alive for end-users, enabling everyone to see the story behind the code.

Project Repository URL: https://github.com/cxy012/TrendTrackr

Project Video

🪙Finesse.ai

Our inspiration stems from the growing complexity of personal finance management. Many individuals struggle to track their expenses, income, and transfers, especially when it comes to synchronizing data across different devices and platforms. We aim to harness AI technology to create an intelligent, personalized solution that empowers users to effortlessly manage all aspects of their finances and make informed financial decisions.

The AI Financial Assistant provides users with a smart and convenient way to manage their finances. Users can log expenses, income, and transfers, adding details such as payment locations and timestamps, while also supporting various subscription plans. The AI analyzes users’ financial patterns to offer personalized insights and recommendations. Additionally, the app features identity verification, notification reminders, and secure data synchronization across multiple devices, ensuring that users can safely and easily access and manage their financial information.

Project Repository URL: https://github.com/Ada6-6/billServer and https://github.com/Ada6-6/billsBot

Project Video

🪙SmartAInventory

The world of retail is a delicate balancing act—maintaining optimal inventory without overspending or running out of stock. Inspired by the potential of data-driven decision-making, we set out to bridge this gap using cutting-edge AI and Microsoft technologies. Our vision was to empower businesses to transform raw data into strategic assets, ensuring efficiency and maximizing profitability.

Throughout this journey, we deepened our understanding of the capabilities of Microsoft Fabric, Azure SQL, and AI integration. We learned that sometimes simplicity, like using batch processing for strategic insights, can deliver more value than overcomplicated real-time analytics. We discovered the power of aligning technology with business needs and how to transform complex data into intuitive, actionable insights.

Project Repository URL: https://github.com/mneang/SmartAInventory

Project Video

🪙Prompt ReAct: Real-Time Voice Intelligence

As conversational AI becomes increasingly central to business operations, the need to optimize AI prompts and responses in real time has grown exponentially. However, existing tools for prompt engineers often fall short, especially when it comes to integrating voice interactions with real-time analytics. This gap inspired us to develop Prompt ReAct—a platform designed to empower prompt engineers with actionable insights directly within Power BI Embedded.

Our vision was simple: leverage Microsoft Fabric and Azure AI Services to create an environment where prompt engineers can refine AI interactions as they happen. We wanted to make Power BI Embedded dashboards not just a reporting tool, but a dynamic space for optimizing AI conversations in real-time.

Building Prompt ReAct deepened our understanding of integrating Microsoft’s ecosystem to unlock the potential of conversational AI:

• Power BI Embedded within Microsoft Fabric: Embedding analytics directly into Power BI allowed us to transform dashboards into interactive hubs for analyzing and optimizing AI prompts in real-time.
• Azure OpenAI (GPT-4o): By integrating GPT-4o’s capabilities, we enhanced our ability to process both voice and text data, providing immediate insights for prompt adjustments.
• Azure AI Search Services: Implementing voice analytics with speech-to-text and text-to-speech capabilities enabled real-time interactions that are analyzed within Power BI.
• Feedback Loop with Cosmos DB & Azure Cognitive Search: Leveraging data stored in Azure Cosmos DB and indexed by Azure Cognitive Search, we created a continuous cycle of AI prompt optimization based on real-time user feedback.

Project Repository URL: https://github.com/mneang/SmartAInventory

Project Video

Congratulations 💜

Congratulations and thank you all who joined us for the Microsoft Fabric and AI Learning Hackathon! Whether you snagged a prize or not, you’re all champions in our eyes! 🎉 You took the leap to learn a new skill and made a dedicated effort toward it, kudos to you! 🌟👏

A special thank you to our judges for all their help with evaluating the projects and selecting the top entries: Jasmine Greenaway, Ismael Mejía Useche, Alvaro Videla Godoy, Josh Ndemenge, Someleze Diko, Swetha Mannepalli, Miwa Monji, Muazma Zahid, Santhosh Reddy Vootukuri, and Siri Varma Vegiraju.

We’d love to hear what you think in the comments. Let us know if there is a particular solution that stood out to you how you think it could provide value to end-users! If you would like to continue the discussion of this event, check out the Hack Together discussion board on the Microsoft Fabric Community site!