About-Me

Welcome to my Github Profile .

Manjinder Singh

I am a seasoned Data Scientist with a proven track record who has worked on several data science problems including Time Series, OCR Text Detection, and Object Detection using CNN and NLP.

While building projects, my primary focus is on creating a Production Ready Model. While working on any project, I divide the whole process into 3 processes:

Research Phase - This is the phase where I analyze, visualize, transform, and extract features from my data in order to come up with an optimal model along with all the essential features.
Packaging Phase - In this phase, I compile all of my analysis done in the research phase and build a CI/CD pipeline that can be dockerized and deployed on the cloud.
Deployment Phase - This is the most exciting phase for me. In this phase, I deploy the model on the AWS Cloud and then try the whole flow using completely new data in order to see if the code breaks or not.

You will see all of my latest repositories designed in this fashion.

A major part of my repositories are focused on the below problems:

Object Detection using Yolo
Sentiment Analysis using Gensim(NLP Package)
AWS Cloud Deployment
Time Series Forecast using ARIMA and Prophet
OCT Detection using Detectron and Tesseract
Sentiment Analysis with BERT using HuggingFace

Manjinder Singh's Projects

car-sales-prediction

Utilizing Neural Networks, predict car purchase amounts accurately. With cleaned data, train the model and assess its performance using MSE & R². Unveil insights into customer spending patterns effortlessly

deploying-machine-learning-models

Code for the online course "Deployment of Machine Learning Models"

end-to-end-ml-project

An End-to-End ML Project predicting student math scores leveraging features like race, gender, parental education, etc. Aimed at deploying a production-ready model on AWS, Azure, etc. Includes EDA, feature engineering, model creation with Random Forest Regressor, achieving 0.87 Adjusted R-Square & 4.3 MSE.

garbage-category-segregation-using-yolo

In 2022, Langara College researchers collaborated with the British Columbia government to pioneer Automatic Rubbish Segregation. With a manually collected dataset of 6000 images, we balanced organic and non-organic waste for training. After experimenting with various CNN models, we opted for Yolo V8 for its optimal balance of accuracy and speed.

manjinder22fzk

Welcome to my Github Profile .

mean-auth-app

Complete MEAN stack app with authentication

mern

Mern Poolcarz

ml-deployment-aws

It is my solution to Kaggle Competition - https://www.kaggle.com/c/house-prices-advanced-regression-techniques/data

sentiment-analysis-using-bert

Explore sentiment analysis with BERT through a detailed tutorial. Learn setup with TensorFlow, numpy, and Hugging Face libraries. Dive into data preprocessing, model training, optimization, and evaluation for practical insights. Perfect for understanding and implementing sentiment analysis in NLP.

Recommend Projects