Projects
Object Detection and Tracking Nov ’23-Dec ‘23
Developed an algorithm to track and count person in a video by calculating
Euclidian distance between centroids of new and previous model using YOLOv8 and finetuning with custom dataset for better results.
UMass 3D Model Sep-Dec ‘23
A 3D visualization of University of Massachusetts Boston building models using blender and rendering them using three.js with floor maps for each floor for better navigation.
It also includes 3D visualization of university from inside.
Predicting Hand Gestures Sep-Nov ‘23
Combining multivariate data of 25 dimensions and 51 timesteps into one, then creating atomic
units by using clustering for each sample then using different models to train and predict 6 various hand gestures on deck of aircraft carrier.
FaceRank Website June ’23-Aug ‘23
Developed a website MERN stack to collect data of ranking 3 best and 3 worst face images
from a set of images of different users and then used this data to make a classifier to perform this action on its own to improve recognition.
Real-State Investment Website Feb ’23-May ‘23
Developed a website using Django stack to show properties available for buying on
location and implement a calculator which calculated cash on cash ROI and cashflow along with other relevant details from given data.
Object Detection Dec ’22-Jan ‘23
Developing a model by combining bounding boxes estimation (Regression) and Class
prediction of different targets (Classification) and detect multiple objects in an image using it.
Forecast Cab Booking, Oct-Dec ‘22
To combine historical usage patterns with publicly available data sources, such as
weather data, to predict if most people will book a cab in a city. Based on this, the company could determine whether to deploy
more or fewer cabs on that specific day. Analysing the dataset and used Feature Engineering to select features by finding
patterns in the data. Fitting various models and manually selecting best parameter using algorithms to perform predictions.
Auto Insurance Industry, Aug ‘22
To predict whether an owner will initiate an auto insurance claim in the next year.
Involved feature Engineering, dealing with unbalanced data, determining which model fits best with best hyper parameters.
Handwritten Character Recognition, Jul ‘22
The aim of this project is to automatically convert handwritten text into
machine encoded text using deep learning, natural language processing, feature extraction and CTC decode layer.
Recognition of plates in moving cars, Apr ‘22
The objective is to detect Moving Cars and License plate in a video file
using OpenCV and object detection using pretrained models.
Recognition of Facemasks, Mar ’22
The goal is to create a Deep Learning model to detect in real-time whether a person
is wearing a face mask or not by implementing and using different pretrained models like MobileNetV2.
Property Sale, Oct ’21
Created a system for purchasing and selling property using Ethereum Solidity
Ballot Voting System, September 2021
Created a voting algorithm using Ethereum Solidity
Redesign Marvel Site, May ‘21
Created a website using Html, CSS, JavaScript and jQuery

Browsing and Purchasing Cars, Apr ‘21
Created a website using Html, CSS, JavaScript and jQuery

Housing Prices Competition, Sep ’20
For predicting the final price of each home and to overcome the problems of
missing data, using advanced regression techniques like random forest and gradient boost
Covid Trend, Aug ‘20
Analysed data of different countries and predicted trends in their patient's rate
Translation of DNA to Protein, Aug ‘20
Separated DNA sequence to amino acid code and then to protein.
Reading Digital Books, Jul ‘20
Decoded encrypted books to find its author, language, title, unique words.
Bird Tracking, Jul ‘20
Designed a bird tracking system with data of their positions in latitude and longitude and
plots them to see the trend. The project predicted the routes various birds took during different time of year.
Connect Dot Game, Jul ‘20
Built a reasonably intelligent agent with the correct algorithm. The agent assumes that
its opponent plays optimally (with respect to the heuristic and using a game tree of limited depth).
Cleaning for Date, Jun ‘20
Worked with messy medical data, used regular expression to extract relevant information
from the data (identify all different date variants) and properly normalized and sorted dates.
Creating a Spelling Recommender, Jun ‘20-
Created different spelling recommenders (based on shortest distance),
that takes a misspelled word and recommends a correctly word using Natural Language Toolkit.