RezScore Jobs
Score the Best

Kaustubh sivalenka
Stony Brook, New York
Stony Brook University, Stony Brook, New York
Graduation: Spring (May) 2020
Master of Science, Computer and Information Sciences
GPA 3.5/4.0
Courses Computer Vision, Big Data Analytics, Natural Language Processing,
Database Systems, Analysis of Algorithms, Probability and Statistics
Mahindra Ecole Centrale, Hyderabad, India
August 2014 - August 2018
Bachelor of Technology (B.Tech), Computer Science and Engineering
GPA - 8.80/10.0
Dept. of Bio-Medical Informatics, Stony Brook University
January 2019 - Present
Graduate Research Assistant
• Developed a gradient-based adaptive sampling prediction scheme for precise discovery of tumor
regions in tissue images (WSI).
• Currently, integrating this prediction scheme with data pipelines based on InceptionV3 and
VGG-19 networks.
UNICEF - Office of Innovation, New York
June 2019 - September 2019
Machine Learning Intern - Satellite Imagery Analysis
• Land Cover Mapping - Developed a U-Net based segmentation analysis pipeline to identify
land resource usage from satellite imagery. Recorded validation f1-score of 88.53%.
• Poverty mapping - Developed a transfer learning pipeline using VGG-16 to assess poverty
distribution from satellite imagery. Recorded Spearman Correlation of 0.47.
ISRO, National Remote Sensing Center, India
September 2017 - May 2018
Research Assistant - Predicting Primary Crop Footprint from Satellite Images.
• Developed a ResNet (Convolutional Neural Network) based feature extractor and recorded a
prediction accuracy of 97.43%. This model can also be adapted to novel satellite scenes.
Kohli Center on Intelligent Systems, India
May 2017 - July 2017
Research Intern - Dynamic Winner Prediction in Twenty20 cricket
• Investigated this problem using supervised learning techniques (KNN, SVM, RF, XGBoost).
• Random Forest Classification yielded the highest prediction accuracy of 75.68%.
Programming/ Scripting Languages
Python, Java, C, MATLAB, JavaScript, Bash
Tensorflow, PyTorch, Spark, MySQL, AWS, Hadoop, D3.js
Data Visualization Dashboard: Visual Analysis of Security Incidents
Data visualization, Time Series analysis - (D3.js, Flask)
• Built a web framework to analyze and visualize various attributes of security incidents.
• ARMA based time-series forecasting also informs the possible trend of future incidents.
Crop Yield Prediction - Using Deep Learning
Big Data Analytics, Machine Learning - (Spark, Tensorflow)
• Developed a data pipeline to estimate future crop yield from satellite data (dataset size: 62 GB).
• Data pre-processing to extract histogram features from satellite data was performed using Spark.
• A stacked-two layer LSTM network was used for predicting crop yield from the histogram features.
1. Sasank Viswanadha, Kaustubh Sivalenka, Madan Gopal Jhawar, Vikram Pudi. ”Dynamic Winner
Prediction in Twenty20 Cricket: Based on Relative Team Strengths.” Workshop of Machine Learning in
Sports Analytics 2017 at the European Conference for Machine Learning.
• Recipient: Platinum and Team Leader Award - Apollo AI and IEEE Scholarship for Machine Learning