https://github.com/harshit-saraswat/myprojectsportfolio

This repository serves as my personal projects portfolio.
https://github.com/harshit-saraswat/myprojectsportfolio

portfolio portfolio-page project-management projects projects-list

Last synced: 22 days ago
JSON representation

This repository serves as my personal projects portfolio.

• Host: GitHub
• URL: https://github.com/harshit-saraswat/myprojectsportfolio
• Owner: harshit-saraswat
• Created: 2020-08-23T21:22:10.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2020-08-30T09:30:17.000Z (over 4 years ago)
• Last Synced: 2024-11-13T11:35:56.842Z (3 months ago)
• Topics: portfolio, portfolio-page, project-management, projects, projects-list
• Homepage: https://harshit-saraswat.github.io/css-personal-cv/
• Size: 10.7 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: Readme.md

Awesome Lists containing this project

README

        
# MyProjectsPortfolio

This repository serves as my personal projects portfolio. Here you'll find projects from different areas I've done so far.

Also, some of the projects which I've worked upon won't have a repository or live link because my Non-Disclosure with the organizations/clients I worked with, binds me from sharing them publically.

# Table of Contents

[1. Ongoing Projects](#Ongoing-Projects)


[2. AI-ML Projects](#AI-ML-Projects)


    [2.1. Computer Vision Projects](#computer-vision-projects)


    [2.2. NLP-Projects](#nlp-projects)


    [2.3. Machine Learning/Data Science Projects](#machine-learningdata-science-projects)


    [2.4. Forecasting/Analytics Projects](#forecastinganalytics-projects)


[3. Web Development Projects](#Web-Development-Projects)


[4. Android App Development Projects](#Android-App-Development-Projects)


[5. Other](#Other)


    [5.1. Blogs](#blogs)


    [5.2. Hackathon Mentor Projects](#projects-as-a-mentor-in-hackathons)


    [5.3. Minor Utilities](#minor-utilities)


# Ongoing Projects

### FaceRec:

***Role:*** Full Stack Developer (Website, Database and AI Modules)

***Technology Stack:*** HTML5/CSS3/Bootstrap4,JavaScript-ES6, Python, Flask, MySQL, TensorFlow, OpenCV, Other Python Libraries

***Description:***

FaceRec is a web based analytics dashboard project which provides insights about usage of our custom built Face AI Package and APIs. The project is built to easily manage and gain insights about our tool by understanding client usage behaviors and derived analytics. The insights drawn are on the basis of location, time, module of usage, etc. The AI Module provides features for Face Detection, Recognition, Face Login, Face Verification, Human Face Checks for optimal images, Facial Landmarks detection, etc.

**Project Link:**  *****NDA***



### ExamProctor:

***Role:*** AI Engineer (Computer Vision and Analytics Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, Flask, Pandas, NumPy and Other Python Libraries, HTML5/CSS3/Bootstrap4, Javascript-ES6, C#, .NET, Angular, etc.

***Description:***

ExamProctor is an AI based examination proctoring tool which helps register students for an online examination and monitors their actions during the time of examination using computer vision and nlp techniques. It provides features for candidate/proctor registration and monitor them using Face Detection, Face Recognition, Speech Detection, Speech To text Conversion, NLP, Object Detection, Locked/Safe Browsing, OCR, etc. to check if a candidate is cheating or not and if found cheating then an appropriate flag is raised.

At the end of examination, the admins/proctors can see  the entire candidate list with analytics of various flags raised and generates a Candidate Credibility Score which can be used by admins to cancel or accept the candidate's exam.

**Project Link:**  *****NDA***



# AI-ML Projects

[1. Computer Vision Projects](#computer-vision-projects)


[2. NLP-Projects](#nlp-projects)


[3. Machine Learning/Data Science Projects](#machine-learningdata-science-projects)


[4. Forecasting/Analytics Projects](#forecastinganalytics-projects)


## Computer Vision Projects

### Face Mask Detection:

***Role:*** AI Engineer (Computer Vision and Analytics Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV and Other Python Libraries.

***Description:***

This is a Computer Vision Project which served as one of the modules for my Final Year Project- *Implementation of Face Detection and Object Tracking Techniques to Develop Covid19 Utilities for Face Mask Detection and Social Distancing Violations Detection*. The project aims to detect if a person is wearing a mask or not in the camera feed. It is based on techniques such as:- Image Processing, Face Detection, Image Classification, Transfer Learning, etc.

**Project Link:**  [Face-Mask-Detection](https://github.com/harshit-saraswat/Face-Mask-Detection)



### Social Distancing Violations Detection:

***Role:*** AI Engineer (Computer Vision and Analytics Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV and Other Python Libraries.

***Description:***

This is a Computer Vision Project which served as one of the modules for my Final Year Project- *Implementation of Face Detection and Object Tracking Techniques to Develop Covid19 Utilities for Face Mask Detection and Social Distancing Violations Detection*. This repo contains code for a covid19 utility- social distancing detection using Tiny Yolo Model to be able to run on smaller GPU systems too. It is based on techniques such as:- Image Processing, Object Detection and Tracking, Transfer Learning, Euclidean Distances and Combinations, etc.

**Project Link:**  [Social-Distancing-Detection](https://github.com/harshit-saraswat/Social-Distancing-Detection)



### SnipNow:

***Role:*** AI Engineer (Computer Vision and API Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, TesseractOCR, C#, .NET, HTML5/CSS3/Bootstrap4, Javascript, Angular, etc.

***Description:***

A Computer Vision and OCR based Image Cutting/Cropping tool used to crop/snip out regions from word/pdf documents (Question Papers) and then use computer vision and ocr techniques to enhance the quality of snipped regions, save any excess white spaces by cropping the image to fit content, auto align text, split options for MCQs, remove labels for questions and options, etc.

Helped in reducing the manual work of employees by streamlining the complete process and easy management provisions. Workload decreased by 15% which resulted in faster deliveries.

**Project Link:**  *****NDA***



### Facify:

***Role:*** AI Engineer (Computer Vision and API Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, DLIB, Face_Recognition, Flask, MS SQL Server, C#, .NET, HTML/CSS/Bootstrap, JavaScript, VueJS, etc.

***Description:***

Highly Efficient and Strict Face Recognition System to find duplicate/unacceptable images in an image dataset of thousands of images to monitor exam registrations. The requirement was to recognize faces with only 1 image sample per user. Helped reduce the False/Unacceptable Registrations by 80%.

Added to that, there were a series of checks each image had to pass through before being accepted inorder to confirm the candidate's validity.

**Project Link:**  *****NDA***



### Office Bus:

***Role:*** AI and Backend Engineer (Computer Vision, Analytics, Database and API Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, DLIB, Face_Recognition, Flask, MySQL, Android Studio (XML+Java), etc.

***Description:***

This project has 2 separate modules:

1. Admin Panel for accessing and managing employees and their attendances (Android App)

2. Employee Attendance App for Tab/Mobile Devices to mark attendance using Face Recognition.

The project helps in using contact-less and more secure and error free way of marking attendances/checking in the offices/schools based on face recognition techniques.

**Project Link:**  *****NDA***



### Drive Awake:

***Role:*** AI Engineer (Computer Vision Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, DLIB and Other Python Libraries.

***Description:***

The project served as my 2nd year final project. It aims to detect a driver face and then detect the facial landmarks of that face. We then use that information to detect if a driver is drowsy/sleepy while driving or not. If the driver is found sleepy then an alert sound is raised inorder to wake him/her up.

**Project Link:**  [Drive-Awake](https://github.com/harshit-saraswat/Drive-Awake)



### Come Talk:

***Role:*** AI Engineer (Computer Vision Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras, OpenCV, DLIB and Other Python Libraries.

***Description:***

This project is an American Sign Language Recognizer which detects hand region using countour extraction and detects the hand signs made by passing those through a CNN Model to check what letter was produced and print on the screen. It intended to serve as an aid to deaf and dumb people to help them communicate with the people around them with ease.

**Project Link:**  [Come-Talk](https://github.com/harshit-saraswat/Come-Talk)



### Implementation of Transfer Learning to detect Pneumonia in Chest X-rays:

***Role:*** AI Engineer (Computer Vision Backend)

***Technology Stack:*** Python, SKLearn, TensorFlow, Keras and Other Python Libraries.

***Description:***

The project aimed to use Transfer Learning Technique to detect pneumonia from chest X-ray Images. The model used was VGG19 for this problem statement. Achieved an accuracy of over 92% with an over all precision of 87%.

**Project Link:**  [Pneumonia-Prediction](https://github.com/harshit-saraswat/Pneumonia-Prediction-using-Transfer-Learning-over-vgg19)



## NLP Projects

### Text Spam Classification:

***Role:*** AI Engineer (NLP Backend)

***Technology Stack:*** Python, SKLearn, NLTK and Other Python Libraries.

***Description:***

The project uses basic NLP Techniques such as Porter Stemming, WordNet Lemmatization, Tokenization, Stopwords Removal, Bag of Words Model and TF-IDF Model. The end output is the prediction of a message text as spam or not.

The dataset is highly imbalanced so the precision is low but can be solved via upsampling/downsampling it.

**Project Link:**  [Text-Spam-Classification](https://github.com/harshit-saraswat/NLP-Spam-Classification)



### Dr. Digi:

***Role:*** AI Engineer (Chatbot Designer and Android App Developer)

***Technology Stack:*** Python, SKLearn, DialogFlow, Flask, Android Studio(XML+Java), AR(Unity, C#), etc.

***Description:***

Dr. Digi serves as a personal medical assistant built using space age technologies of Chatbots, Android Apps, AR, etc. It's main purpose is to digitize healthcare as it helps patients find nearest hospitals, guide their ways in a hospital using AR based Indoor Maps, book appointments, manage patient history, etc. 

**Project Link:**  *****NDA***



## Machine Learning/Data Science Projects

### Employee Attrition Rate Prediction:

***Role:*** Participant (Complete ML Workflow)

***Technology Stack:*** Python, SKLearn, Pandas, PyCaret, etc.

***Description:***

Employee Attrition Rate Prediction is a HackerEarth Machine Learning Challenge held in July, 2020. 

The problem statement was to detect what probability is there that an employee will leave you. 

I had to work on complete project starting from EDA Phase to Feature Engineering, Data Preparation, Model Building, Hyperparameter Tuning, etc.

Secured a Top 3% rank in this competition.

**Project Link:**  [Employee-Attrition-Rate-Prediction](https://github.com/harshit-saraswat/Employee-Attrition-Rate-Prediction)



### Machine Learning Bootcamp:

***Role:*** Student and Developer (Complete ML Workflow)

***Technology Stack:*** Python, SKLearn, Pandas, Matplotlib, Seaborn, NumPy, etc.

***Description:***

Worked on multiple small ML Projects in the following domains while studing ML and Data Science as an industrial intern:

1. Regression

    

    a. Linear Regression


    b. Multiple Linear Regression


    c. Polynomial Regression


    d. XGBoost Regression


    e. RandomForest Regression


    f. Ridge Regression


    g. Bayesian Ridge Regression


    h. Lasso Regression


2. Classification

    a. Logistic Regression


    b. Naive Bayes Classifier


    c. Support Vector Machine


    d. Decision Tree


    e. RandomForest


    f. XGBoost


    e. RandomForest


    f. K Nearest Neighbor


3. Clustering

    a. K Means Clustering


    b. Hierarchical Clustering


4. Others

    a. Apriori Algorithm


    b. PCA


    c. NLTK




## Forecasting/Analytics Projects

### HealthCalc:

***Role:*** AI Engineer (AI and Analytics Backend)

***Technology Stack:*** Python, SKLearn, Pandas, NumPy, StatsModels, Keras, TensorFlow, C#, .NET, Angular, HTML, CSS, JavaScript, etc.

***Description:***

HealthCalc is a hospital analytics dashboard project which helped hospitals to forecast future patient flow for various departments such as Cardio, Physio, General, Emergency, etc. and based on that prepare and manage Doctors, Nurses, Sanitation Staff, Ambulances, etc.

The analytics backend predicted all those things on a daily, weekly, fortnightly and monthly basis.

Helped Reduce Hospital Resources Overutilisation by 13% and Underutilisation by 17%.

Overall RMSE was 2.8 Patients With Mean of 130 patients on a daily basis. 

**Project Link:**  *****NDA***



# Web Development Projects

### Simple Flask Blog App:

***Role:*** Full Stack Developer

***Technology Stack:*** HTML, CSS, Bootstrap, Flask, SQLAlchemy, Python, etc.

***Description:***

Simple Flask Blog app is a simple blogging app with user management, authentication module, posts management and other features for a simple blogging website using Flask and Jinja 2 Logic. It also has a Database Connectivity using SQLLite DB. The Project follows a MVC Architecture.

**Project Link:**  [Simple-Blog-w-Flask](https://github.com/harshit-saraswat/Simple-Blog-w-Flask)



### CSS Personal CV:

***Role:*** Web Developer

***Technology Stack:*** HTML, CSS, Bootstrap.

***Description:***

Simple website which serves as my web CV hosted on GitHub Pages. It was one of my first Web Development Bootcamp Projects.

**Project Link:**  [css-personal-cv](https://harshit-saraswat.github.io/css-personal-cv/)



# Android App Development Projects

### eOPD:

***Role:*** Android App Developer

***Technology Stack:*** Android Studio (XML+Java), Xampp Server(MySQL DB+PHP Scripts), etc.

***Description:***

A simple android app to digitally preserve patient records in a secure manner and save them from the hassles of maintaining and handling large amount of medical documents by storing Patient History as a QR Code and sharing it via Image. Doctors registered with the app can have access to write or edit to that data only and no one else.

**Project Link:**  [eOPD](https://github.com/harshit-saraswat/eOPD)



### Tribe-Mail:

***Role:*** Android App Developer

***Technology Stack:*** Android Studio (XML+Java), Xampp Server(MySQL DB+PHP Scripts), etc.

***Description:***

An android application made for SIH 2018 which serves as the one step solution for bringing the government closer to the STs and provide all government services on a single app. The project held an honorary position at SIH 2018. 

**Project Link:**  [Tribe-Mail](https://github.com/harshit-saraswat/Tribe-Mail)



# Other

[1. Blogs](#blogs)


[2. Hackathon Mentor Projects](#projects-as-a-mentor-in-hackathons)


[3. Minor Utilities](#minor-utilities)


## Blogs:

### Face-Recognition-based-Image-Separator:

***Role:*** AI Engineer(Computer Vision and Analytics Backend) and Author

***Technology Stack:*** Python, OpenCV, DLIB, Face_recognition and Other Python Libraries.

***Description:***

This is a small fun blog-project which uses face recognition techniques to separate images from a large dataset into images of different people according to faces.

**Project Link:**  [Face-Recognition-based-Image-Separator](https://medium.com/analytics-vidhya/face-recognition-based-image-separator-408681f2360d)



## Projects as a Mentor in Hackathons:

### SaferWherever:

***Role:*** Mentor

***Technology Stack:*** Python, OpenCV, DLIB, Face_recognition, Streamlit, Flask, TensorFlow and Other Python Libraries.

***Description:***

The aim of this project is to develop a solution based on the use of AI for a post COVID-19 world as part of the #BuildwithAI : Emergence! Hackathon. The Project won **1st Runner Up Position at #BuildWithAI Global Hackathon** with over **5000 Participants globally**. The project processes camera feeds in malls to check for social distancing violations, face mask violations, density in areas and then visualize them on a Dashboard in the form of analytics and hotspot maps.

**Project Link:**  [SaferWherever](https://github.com/tugrulguner/SaferWherever-BuildwithAI-Emergence) or [SaferWherever-Heroku](https://a-team-mall-monitor.herokuapp.com/) 



### CovidPredictor:

***Role:*** Mentor

***Technology Stack:*** Python, Pandas, NumPy, Matplotlib, Seaborn, Statsmodels.

***Description:***

The aim of this project was to use time series forecasting and statistical techniques to predict Covid19 Cases between July 25, 2020 to August 17, 2020.

**Project Link:** [Covid19PredictiveModelling](https://github.com/lebro-23/CovidHackathonPredictiveModelling)



### CovidChatbot:

***Role:*** Mentor

***Technology Stack:*** Python, Pandas, NumPy, Matplotlib, Seaborn, Statsmodels, Flask, Rasa, NLTK, etc.

***Description:***

CovidChatbot is a simple chatbot app which analyses tweets from people with respect to Covid19 and produces users with insights on the abusive or false news in an area along with regular covid 19 question answers.



## Minor Utilities:

### Scraping Bots:

***Role:*** Script Writer (Python Script)

***Technology Stack:*** Python, Selenium, Beautiful Soup, Requests, Pandas.

***Description:***

Wrote a number of scripts to scrape data of internet for data collection or resource gathering from multiple websites using web automated crawlers and exported data in an excel/csv format.