Software Engineer interested in building highly available, resilient, fault tolerant and scalable systems

Harinadh Appidi  • Peerlist

	Predict if patient will be readmitted within 15 days of discharge.
	Implemented word embedding technique as a dimensionality reduction for categorical variables.
	Achieved 1st position in Kaggle in-class competition.

Hospital Readmission Prediction

> Webscrapping of incident data from Norman Police department website.
> Data extraction from pdf files using PYPDF2 package in Python
> Data cleaning and transformation to tabularized format for database insertion.
> Created and Populated data into sqlite database.
> Summarized the incident data and shown to user through command line.

Incidents Summarization

	Implemented RDBMS for a Job scheduling application in Azure.
	Presentation and Business layers are written in Java and database layer in SQL.

Job Shop Accounting System

Predict heart disease based on certain health condition parameter values.

Heart disease prediction

   Designed and developed neural network architectures to perform pixel level classification using satellite images.
   Achieved an average test accuracy of 92% and F1-score of 0.9.

Semantic Labeling of Chesapeake Satellite Images

> Redaction of sensitive information in '.txt' and '.md' files.
> Sentence tokenization of text data.
> Entity recognition to identify Person, GPE and Organization using NLTK and Spacy.
> Rule based techniques to include outlier cases where  enitites are not identified using Spacy and NLTK.

Text Redactor

>Top 5 Actors are predicted in the Redacted IMDB movie review text files.
> Performed Feature Engineering on unredacted training files to identify the best suitable features for the use case.
> Trained Random Forest classifier model on unredacted text files with feature vectors.
> Prediction on redacted files to predict top 5 actors.

Text UnRedactor

	Predict total expected sales revenue from customers across all of their online visits.
	Implemented OLS, MARS, ElasticNet, SVM & Gradient Boosting Machines and achieved 1st position in Kaggle  in-class competition.