Designed an algorithm with 50+ pipelines to process the entire census process for Canada Census (HR) Automation project which reduced the manual processing time from 1-2 days to 2-3 Minutes with well over 100 per cent accuracy.
Worked on Natural Language Processing project, extracted data from the CVs and Resume and Clustered them into different categories.
The aim of this system is to automate helmet violation detection to reduce manual human involvement and to deduct the extensive
utilization of resources for the government.
The ultimate model is fine-tuned and integrated with three distict AI models together including the Computer Vision models to fetch the coordinates of the whole
vehicle and the identification number, and an OCR model to extract the textual information from the provided coordinates.
The project aimed at creating legitimate and fluent questions based on the provided passage and the intended
answers.
The system provides features such as automatically extracting text from a photo for evaluation purposes,
enabling the types of questions to be selected and producing a number of questions including MCQs, Boolean, short
questions.
A Web-Based Machine Learning Pattern Recognition and Sentiment Analysis system to recognize handwritten
images of the feedback forms and perform the image-to-text conversion.
The system was implemented with 93.13%
accuracy on over 1,28,000 customer reviews in the training dataset and a validation set of 32,000 customer reviews.
The Classification goal was to predict if the client would or would not subscribe to the bank term deposit in the
future by applying Machine-Learning Algorithms.
The project is comprised of various crucial tasks for a superior ML
model including Data Pre-processing, Exploratory Data Analysis and Examination of Evaluation metrics for
performance measurement.
ATS is an Android application which enables the user to find the stolen phone, it operates in the context without the
person in connection.
The application works by storing the IMSI number and continuously checking for a SIM
Ejection. Upon the detection of a suspicious SIM Ejection, it will ultimately send an SMS to the alternative number
with an exact location.
This project has several other features like Data-Wipeout, Virtual control over a device, the
option to hide an app, etc.
Notifiey is a web application that provides an interactive platform where faculty members can leave the updates,
events, public announcements, etc.
This system helped students to get notified about the latest updates on the
upcoming events in the college while making it easy to share for the faculty members.
Trained and Tested a Model on 1.6 million sentiments using various Machine Learning and Deep Learning Algorithms including Support Vector Machine (SVM), Multinomial Naive Bayes Classifier, Bidirectional Encoder Representations from Transformers, and Long Short-Term Memory.