Project

This project was part of MetLife Asia’s AI Hackathon 2023, where I led a team of 7 members and we became one of the top 15 finalists. Our aim was to fully automate the Claims Submission Process using the Azure Cognitive Services and Word2vec embedding. This solution would eliminate the risk of human errors and AI document checker would predict the submission validity, resulting in easier decision-making capability for the adjudicators.

Solution Summary:

• Verification of Documents using AI Model during the Submission Stage.

• Mandatory Field Data Entries using Optical Character Recognition (OCR).

Fully Automated Claims Process that will remove Human Dependency.

• Benefits/Impact of Proposed Solution:

• An automated solution that lessens manual intervention and reduces customer waiting time

• No more human dependency for data entry, hence reducing cost

• Lesser number of errors due to incorrect data entry

• Reduced decision-making time for Adjudication, hence improved claims processing time

This project was part of my undergraduate Integrated Design Project (IDP) at Military Institute of Science and Technology (MIST) which was also funded by the MIST Computer Science Department. We chose to work on the then-pressing issue - COVID-19. To replace the traditional temperature measuring method, we developed an affordable and human-contactless temperature sensing and management system. With this, the computer could detect a human’s face, read temperature directly from his/her forehead, and generate alerts if the recorded temperature is above the defined threshold, refraining anyone with potential COVID-19 symptoms from entering the University premises. This system would also automatically capture students’ attendance, reducing the manual effort for teachers.

Read more here.

This project focused on the handloom sector of Bangladesh (BD). Handloom being one of the traditional cottage industries of BD, comes next to be the second largest source of rural income employment after agriculture. Unfortunately, most of the handloom factories of our country are in a dire state today due to negligence and lack of digital awareness. Loom Arena aimed to revive our handloom sector from this dying situation by designing database for the people working in the handloom sector of a specific company where, they will be able to manage the database of the products they sell and communicate directly with buyers, removing any middle-man’s intervention. 

The aim of the project was to produce a digital system capable of encoding ASCII characters and displaying the output as Morse Code representation of the characters on a screen. Morse Code is a method of transmitting string information into a series of dots[.] and dashes[-] encrypting a message between two entities and so making communication possible between two parties. In this system, we used a hex keypad that takes an input and relays this message to arduino. Finally the morse code equivalent of the input message is displayed on the LCD.