""

Takeaways from our Hack-AI-thon

Menlo Technologies AI Hackfest took place on the 23rd & 24th of December, 2019. Ten teams competed for cash prizes during a focused 24-hour effort. They explored concepts and technologies of Data Science/AI with respect to Azure services, as well as gained awareness on the capabilities and the potential of Data Science to solve business problems. Our teams developed a solid enterprise understanding of the Data Science Life Cycle and the various components that form it. 

We have one goal – make Data Science as easy as 1-2-3 for our Developers and in-turn our Clients. Each team addressed a business problem and came up with outstanding solutions with the help of data models and methodologies. On a broader note, these were the topics the Hackathon was focused on :

  1. Video Indexing
  2. Fraud Detection & Anomaly Detection
  3. Predictive Analytics
  4. Sentiment Analysis

Our Development Teams were super curious to deep dive into AI and uncover noteworthy solutions before the bell rang, signaling the end of the competition. The implementation took over 24 hours with a 2-week prep time to investigate and understand the project.

Hackathon participants

Let’s get a closer look into the projects and their key findings :  

Projects

Technologies Used

Key Findings

Use Cases

Video Indexing

Custom face recognition, sentiment analysis and objects identification from the video.

  • Azure Cognitive Services – Video Indexer
  • Power BI
  • Azure Blob Service​
  • Azure Virtual Machine​
  • JavaScript
  • JQuery
  • Node.js
  • Azure App Services, Registrations & Insights
  • Azure Postgres
  • Use of Cognitive Services, Integration of API’s provided by Azure Video Indexing.
  • Face recognition.
  • Sentiment analysis and keyword extraction on audio content.
  • Speech to text transcription.
  • Object identification in the video content.

Facial recognition and language identification by Law Firms, Crime Departments or Security Firms.

Businesses can compare product reviews with competition using keyword extraction.

Video analysis or emotion detection by Media firms.

Fraud Detection & Anomaly Detection

Identifying anomalies in credit card usage.

  • Azure ML Studio
  • Azure SQL
  • Azure Web Services
  • Azure Event Hubs
  • Azure Stream Analytics
  • Power BI
  • Azure Cognitive Services
  • Analyze and tune the credit card transactions to detect possible fraudulent behavior.
  • Azure ML Studio can be leveraged to quickly kick-start a machine learning project without implementing complex algorithms from scratch.
  • Azure has built-in cognitive services to do anomaly detection.

Banks or Financial Institutions can use it for credit card fraud detection and send alerts to users accordingly.

Healthcare Companies can monitor System Health and conduct Predictive Maintenance.

Meteorological Companies can get accurate Weather Information.

Retail sector can do anomaly detection on sales, inventory, pricing and other quantitative attributes.

Predictive Analytics

“Order Demand” prediction in a specific area to help plan accurate logistics.

  • Azure ML Studio
  • Azure SQL
  • Azure Web Services
  • Azure Trained Models
  • Azure Datasets
  • Azure Blob Service
  • Vue JS
  • Node JS
  • Power BI
  • Azure ML Studio is easy to understand and has many built-in functions, examples, datasets.
  • Code snippets provided for consuming web services in Single/Bulk modes are helpful.
  • Easy to train models for classification, regression, clustering & feature extraction.

Wherever demand-supply exists, eg : Stock Market, Investors

Logistics services like food supply companies

Taxi service companies

E-commerce industry

Sentiment Analysis

 Compare product reviews with competitors to understand what your customers like and dislike about the product/service.

  • Cognitive Services
  • Microsoft Azure Blob
  • Microsoft Azure Machine Learning Studio
  • Azure Sentiment Analysis API
  • Azure Speaker Recognition API
  • Azure Linguistic Analysis API
  • Azure Text Conversion API
  • Azure Text Processor API
  • Azure Key Phrase Extraction API
  • Python
  • Azure NoteBook
  • Power BI
  • Java (SpringBoot)
  • Angular
  • Azure ML Studio is easy to understand and has many built-in functions, examples, datasets.
  • Code snippets provided for consuming web services in Single/Bulk modes are helpful.
  • Easy to train models for classification, regression, clustering & feature extraction.

Wherever demand-supply exists, eg : Stock Market, Investors

Logistics services like food supply companies

Taxi service companies

E-commerce industry

Although these ten teams came up with unique use case solutions, the competition itself cannot be ignored. So let’s give the winners their share of limelight –

Menlo AI hackathon winners

Added Value

  • Team-bonding – organizing and executing hackathons is a good diversion from the regular work routine, makes the workspace more interactive and inspires people to learn something new.
  • Learning – hackathons are for learning. You’ll gain more practical knowledge in one hackathon than you learn in a month of lectures. Our developers love hackathons as it gives them an opportunity to dive carefree into topics, test limits and gain abundant real-time experience.
  • Solving business problems – it turns out that these solutions often times help our clients too. In fact, we have already implemented Sentiment Analysis solutions for our client’s ML projects.

Summary 

With ten teams composed of members with different expertise, interacting on AI technologies to solve business problems, was an invaluable benefit of this Hackfest.

Entire Menlo team

Menlo Technologies  is a Microsoft Gold Application Development & Data Analytics Partner. Learn more about our Data and AI capabilities.

Menlo logo
Menlo Microsoft competency logo

Copyright © 2019 Menlo Technologies. All Rights Reserved.

FOLLOW US FOR LATEST UPDATES