Ostap Orishko's repositories
pyspark-triangle-count
Implementation of the triangle count algorithms without using GraphFrames or GraphX in Spark.
VGG-and-ResNet-Architectures
This is an implementation of a number of famous deep learning architectures from scratch using PyTorch. VGG13 and VGG16, as well as ResNet performances are investigated using CIFAR10 and MNIST datasets.
Advanced-ML-Spec
The code I wrote for a Coursera specialization in Advanced Machine Learning. Each folder only contains graded assignments.
CloudComputingProject
StockHelper queries historic stock data for an existing user and helps make notes for user Stock searches. The project involves Flask, Cassandra, Docker and Chart.js. to build a dynamically generated REST API with hash-based user authentication, serving the application over https.
codewars
Code for kata solutions
Coursera-NLP-Specialization
Code for the NLP Coursera Specialization from deeplearning.ai
Face-Recognition-Implementation
The code for creating a data set of images containing a person's face and then using it to recognize any person in the data set on camera.
Guardian-scrape-wordcloud
A quick little project where I use scrapy to get clean data from the Guardian on title and description of articles on a given day for various topics. I then use this data to extract keywords using Rake and create a word cloud out of keywords to get a feel of what topics made the headlines. The generation of word clouds is performed in gen_wordcloud.ipynb and crawling docs are saved into the spiders folder.
UMLEND-model-deployment
This is a simple LSTM model built on top of IMDB movie review dataset, deployed using Amazon Sagemaker, Lamdba and API Gateway used to obtain live model predictions via a web page.
react-weather-app
Weather App built in React.
ydata-synthetic
Synthetic structured data generators
YouTube-Engagement-Prediction
Data Science team project conducted to explore business value from a dataset containing Trending YouTube videos across a period of time.