Mihir Chakradeo

Zuora

Strategic Solutions Engineer Intern • May, 2018 — Aug, 2018

• Designed, developed, and deployed a Rails Middleware for collecting application metrics like response time, database time, number of resque workers, etc.
• Wrote queries in Grafana for setting up alerts for application failures and anomalies
• Wrote Continuous Queries in InfluxDB to automatically condense old data beyond a retention period for reducing space used
• Technologies: Ruby, Rails, Kubernetes, Prometheus, Telegraf, InfluxDB, Chronograf, Grafana

NLP lab, Stony Brook University

Graduate Student Researcher • Jan, 2018 — Present

• Clinical Trials are hard to find. Searches show a lot of trials which might be irrelevant to the user. The aim of this project is to use Natural Language Processing (NLP) to extract information from the existing trials and ask questions to the user in order to reduce the search space and only return the most relevant trials.
• Stats: Was able to reduce the number of searches of Diabetes related clinical trials in long island from ~80 to ~10
• Technologies: Python, pandas, numpy, tensorflow, matplotlib

SUNY Stony Brook

Master of Science, Computer Science (Data Science Track) • Aug 2017 — Expected Graduation - Dec 2018

University of Pune, India

Bachelor of Engineering, Computer Science Engineering • 2013-2017

ADMM Optimization algorithm in PyTorch

• Mar 2018 — Apr 2018

• Implemented parallelized “Alternating Direction Method of Multipliers” Optimization algorithm
• Tested against sklearn’s LASSO solver on the diabetes dataset, achieved an average speed up of 1.18
• Technologies: Python, PyTorch, numpy, matplotlib

Tags: Deep Learning, Convex Optimization, Machine Learning

Human Action Detection using Deep Learning(CNN)

• Feb 2018 — Feb 2018

• Wrote a Convolutional Neural Network in PyTorch for classifying images and videos of human actions. Used the UCF101 action recognition dataset
• Technologies: Python, pandas, numpy, PyTorch, matplotlib

Cryptocurrency Data Analysis

• Feb 2018 — Mar 2018

• Applied various techniques of Statistical Inferencing and Data Science on the Cryptocurrency Dataset on Kaggle, which contributed towards my final project of Prob Stats for Data Science Grad course
• I worked on finding how Google searches affect crypto prices. Tests applied were - Linear Regression, Wald's Test, KS Test, Multiple Linear Regression to find which crypto was influenced the most
• Technologies: Python, pandas, numpy, sklearn, matplotlib

Google Firebase AppFest - Skin Cancer Detection with a Dermatoscope

• Jun 2017

• Built an android app for detecting skin cancer using a Dermatoscope. Secured second place.
• Trained Inception v2 on images from google search scrapes
• Technologies: Python, pandas, numpy, PyTorch, matplotlib, Android

Languages

Python, C++, Ruby, SQL, Javascript

Libraries

NumPy, SKlearn, Pandas, TensorFlow, PyTorch, Matplotlib

Frameworks

Flask, Rails

Other Technologies

Git, Kubernetes, Docker, InfluxDB, Kafka

Courses

Machine Learning, Convex Optimization, NLP, Computer Vision, Probability & Statistics for Data Science, AI

• Chess, Soccer, Ping Pong
• Music (Keyboard/Piano)
• Gaming