Shrey Arora

Statistics | Machine Learning | Data Engineering

About Me

Your Image

Hi, My name is Shrey Arora. I'm pursuing MS in Statistics and Data Science from University of Arizona, Tucson. I have 3 years of demonstrated work experience in advanced analytics, machine learning and data engineering roles in e-commerce and fin-tech industries. I have a background in Electrical Engineering with bachelors in engineering from Thapar University, India. Other interests and hobbies include hiking, playing football, writing poems, playing the keyboard and producing music.

Education

Thapar University, Punjab India - B.E Electrical Engineering

Aug 2015 – May 2019

Relevant Courses

GPA : 7.62/10


University of Arizona, Tucson AZ - MS Statistics and Data Science

Aug 2022 – May 2024

Relevant Courses

GPA : 3.56/4

Teaching Experience

Graduate Teaching Assistant at department of Physics and Atmospheric Sciences. I taught PHYS 240 Electricity and Magnetism and took lab sessions for two sections

Member of honarary statistics socierty Mu Sigma Rho

Research Interest

Work Experience


Tesla, Fremont CA

Data Engineering Intern

Aug 2023 – Dec 2023


American Express, India

Analyst - Risk & Info Management

Aug 2023 – Dec 2023


Prione (Amazon), India

Business Intelligence Analyst

Sep 2020 – Sep 2021


Happay (CRED), India

Data Scientist

Jun 2019 – Sep 2020

Projects

Artificial Intelligent Pac-man - Search Algorithms | Reinforcement Learning | Optimization | Python

Created an AI enabled Pac-Man agent to optimize decision making in live game . Implemented agents using search algorithms like Depth-First Search, Breadth-First Search, A* ,minimax, expectimax and reinforcement learning (Q-Learning) to make pacman learn through trial and error. Designed better evaluation functions to ensure better in-game performance.


YouTube Comment Multi-Labelling - NLP | Neural Networks | Text Multi Labelling | Hyperparameter tuning

Trained a bidirectional neural network on Youtube comments data that inputs given text and predicts whether any of the given emotions are present: admiration, amusement, gratitude, love, pride, relief and remorse. Used hugging face's Roberta-Base tokeniser and pre-trained model to add embeddings in the model that achieved 84% macro f1 score on test-set.


Huawei Advertisement CTR Prediction - ML Modeling | Resampling |Ensembles |Boosting | Python | VS code | R Studio

Modeled the click-through rates for mobile advertising. Large imbalanced dataset was analyzed, and resampling techniques were deployed with different classification methods to minimize class imbalance. The model was optimized and hypertuned to achieve an 82% f1 score with XGBoost.

Contact

You can reach me at: shreyarora@arizona.edu