EXL™ Hackathon

Overview

The current solution processes backend requests and generates reports in .pdf, .csv, or .pdm file formats. These reports are then uploaded on an on-premises shared file server. Since the solution is now hosted on cloud, a file-server-based model is not the optimum solution for file storage.

Though transitioning file storage from on-premises to cloud can be achieved easily, clients have their choices of cloud provider, which requires customization for every transition. A cloud-agnostic storage solution can address this problem.

Can you build a solution that is agnostic of the cloud provider for storing and retrieving files, i.e. one solution that supports storage and retrieval of files from AWS S3 bucket, Azure Blob storage, or Google Cloud storage.

Tech stack

• AWS
• Azure
• Cloud Storage

Task 

• Minimum
  • Working API-based implementation
    • It should be capable of storing and retrieving files in any cloud provider.
    • It should adhere to the best standard security guidelines.
    • It should support AWS and Azure at least.
  • Intermediate
    • AWS instance
  • Advanced
    • Cloud-provider support can be enhanced to GCP and other cloud providers as well

Submission format 

• Source code in a zip file
• Details about the cloud storage and service provider
• Documents used for security guidelines
• Important files, pdf, resources used and ppt

Resources

• AWS
  • Why AWS
• Azure
  • What is Azure?
• Cloud Agnostic Solution
  • Cloud Agnostic

Overview

Imagine a BOT that takes minimum information and instructions to recommend products to a user. The BOT can be configured by carrying out simple tasks such as uploading a set of questions and answers at a specific URL, which could be an API from the BOT admin. The BOT analyzes the data, makes third-party API calls, and makes product recommendations to the user. Further, the BOT analyses user response and starts recommending other products, like upselling and cross-selling.

The BOT eventually becomes the foundation for a better user experience driven by AI and Machine Learning.

Can you code a BOT that provides valuable information to a user through Q&A, supported by Natural Language Processing (NLP) making it a great training tool to onboard new users or new agents? Built with a two-way integration with the application, the BOT needs to be trained to be specific to their customers and their insurance policy products.

Tech stack

• Web Application
• IOT
• Natural Language Processing
• Machine Learning
• AI

Task

• Minimum
  • Working BOT that is configurable and has a flexible API layer
• Intermediate
  • Intuitive UI/UX  along with BOT
  • Performant and fault-tolerant
  • Scalable design to handle future volumes
• Advanced 
  • Features of NLP to make the experience more interactive
  • Accepting structured as well as unstructured data as inputs during the conversation with the BOT

Submission format 

• Source code in a zip file
• The design used for BOT
• Dataset Used
• Details about the cloud storage and service provider
• Documents used for  security guidelines
• Important files, pdf, resources used and ppt

Resources

• BOT
  • Create a BOT
• Natural Language Processing
  • NLP
• Web Application
  • Web Development

Overview

Appointment scheduling and routing solution helps in reducing manual, time-taking processes such as maintaining data. security, etc. Even tasks such as retrieving past data and reporting details of upcoming tasks can be managed easily.

This solution will also help in tracking locations and addresses. It can be integrated with online maps to improve efficiency, which is not possible in a manual offline process.

Tech stack

• REST API
• Angular
• Scheduling and routing techniques 

Tasks

While building this solution, you need to keep the following points in mind while working on your idea. These points must be integrated with your solution to make the process of appointment scheduling and routing easy and effective for the user.

• Provide a list of addresses to visit and one home address as input
• Visualize these addresses into a Google map
• Define the number of days (Y) across which these addresses need to be distributed (for example, 3 days, 4 days, etc.)
• Define the number of addresses to visit on Day 1, Day 2, up to Day Y 
• The system should suggest locations to visit on each day such that the drive/travel time for the day is optimized to keep the amount of travel time to as minimum as possible.
  • Users can drive between locations selected for the day as well as spends an average of Z hours at each location within the time slot of M to N. 
  • Users should be able to drive back home from the last location in the least possible duration
• The output must have the following details:
  • Day-wise list of locations in sequence
  • Daily route including the drive time from the home address to different locations (in the order as suggested by the system) and back to the home address
    • Configurable input parameters
      • X = Number of addresses (excluding the home address)
      • Y = Number of days
      • Z = Average appointment time (in hours)
      • M = Start time
      • N = End time
      • I = Number of addresses to visit for a particular day
    • Minimum task
      • Build a restful API to schedule and confirm appointments based on the description mentioned.
      • An interface to clearly see the agenda/calendar for the day with the drive time between appointments within an angular application
      • Integration with Outlook/Teams calendar
    • Intermediate task
      • Create a smart solution to rearrange the appointments to reduce drive time
    • Advanced task
      • Integration with Teams or Outlook to reflect the agenda.

Submission format 

• Source code in the zip file
• Details about GitHub file and REST API
• Details about the outlook and team integration.
• Important files, PDF, resources used, and PPT

Resources

• REST API
  • What is REST API?
• Angular Application
  • Angular
• Scheduling and Routing
  • What are scheduling and routing