Machine Learning at UMD

Aditya Patkar  • Peerlist

Finance Head at Bake A Diem - A student established and run bakery venture by the students of Symbiosis School of Culinary Arts.

All the documents including balance sheet, break even analysis, business plan etc. have been made open source and can be found below.

Bake a Diem by SSCA

Analyzed over 300,000 Kickstarter projects spanning over a decade using NumPy, Pandas, Matplotlib and Seaborn to determine overall success rate and other interesting trends.

Data Analysis of Kickstarter Projects

Predicting health conditions using X-Ray
• Built a custom Convolutional Neural Network which predicts occurrence of Covid-19, Viral Pneumonia or Lung Opacity with 94 Percent accuracy.
• Envisioned and implemented a front end login system and a UI to interact with the model using Streamlit, Flask and MongoDB, deployed on cloud.

diagnose.ai

Built a RESNET-like Neural Network in PyTorch to classify images of garbage into 6 different classes with an accuracy of 80%.

Image Classification of Garbage - Deep Learning

A drag and drop style canvas for interacting with powerful Machine Learning algorithms like 
- Face Detection
- Object Detection
- Entity Extraction
- Sentiment Analysis 

technologies used: 
- Python
- Flask
- MongoDB
- SpaCy and NLTK for NLP, YOLO for Computer Vision.
- Heroku for deployment.

Marks: 48/50

MLHub.ai

- A web app to forecast the Air Quality Index (AQI) for the city of Mumbai, India
- Time-series forecasting with 24-hour error (MAPE) revolving around 10%