Data Science Machine Learning

SHIVAM CHOUDHARY

+1-470-***-**** ************@******.*** Atlanta, Georgia

linkedin.com/in/shivam-choudhary-524702105/ github.com/shivamchoudhary2121997 SUMMARY

Experienced Data Scientist with 5+ years of expertise in machine learning, specializing in fraud detection, forecasting models, and time series analysis and worked with organizations like American Express, ReBIT, and Quantiphi. Successfully developed models with significant business impact, such as a GBM model for fraud detection that led to estimated $14 Million in monthly savings based on the test data. Currently pursuing an MS in Analytics

(Data Science) at Georgia Tech, targeting MLE and Data Science Intern roles, with a long-term goal of advancing to senior roles in MLE or data science. EDUCATION

Georgia Institute of Technology Atlanta, US

Master of Science in Analytics (Computational Data Analytics Track) Aug 2024 - Dec 2025 (Expected) Indian Institute of Technology(IIT) Kanpur Kanpur, India B.Tech in Chemical Engineering Jul 2015 - Jun 2019 TECHNICAL SKILLS/RELEVANT COURSES

• Languages, Tools/Libraries and Platforms: Python, R, C++, SQL, Tensorflow, Scikit-Learn, Keras, Numpy, Pandas, Matplotlib, Seaborn, NLTK, Transformers, Pyspark, Git, Excel, GCP, AWS, spacy, Huggingface

• Use Cases: Machine learning, Deep learning, NLP, Time series analysis, Fraud detection

• Relevant Courses: Regression Analysis, Deep Learning (Georgia Tech), Gen AI with LLMs (Coursera) EXPERIENCE

Accertify India Private Limited (Previously a subsidiary of American Express) Gurgaon, India Senior Analyst, Data Science May 2024 - July 2024

• Developed iterations of XGBoost ticketing model using data of 10+ clients, enhancing fraud coverage by 80% over previous models on the test set

American Express India Private Limited Gurgaon, India Senior Analyst, Data Science (Part of the Accertify team) June 2022 - April 2024

• Supervised and collaborated with other analysts in the data preparation and development of GBM model for fraud detection using data of 32 clients resulting in estimated $14M+ monthly savings at the portfolio level. Won Analyst of the Quarter award for this model

• Delivered fraud strategies for major clients in retail and airline industries using Industry level models while achieving review rate reduction of up to 30% and reduction in fraud dollar up to 20%

• Presented and implemented fraud strategies to a major retail client with more than 50 Million transactions per month achieving a 17% reduction in approved fraud dollars in our testing period Reserve Bank Information Technology Private Limited (A subsidiary of RBI) Navi Mumbai, India Data Scientist July 2021 - June 2022

• Developed a short PoC for Question Answering System using a pre-trained BERT model which takes in the document and answers queries based on the document

• Created a PoC for RBI’s statistics and information management department, analyzing borrower default patterns with DBSCAN, k-means, and DTW, achieving a silhouette score of 0.4 Quantiphi Analytics Private Limited Mumbai, India

Machine Learning Engineer May 2019 - July 2021

• Designed a Seq2Seq model outputting one-week sales forecasts, optimized hyperparameter with Sagemaker HPO

• Created a stacking-based model in Python using ARIMA, ETS, LSTM, Prophet, and XGBoost as metalearner, reducing mean absolute percentage error by 20% as compared to production models

• Optimized code calculating forecast with simple statistical formula using vectorization reducing execution time from 20 minutes/store to 1 minute/store

• Mastered ML & DL models during training period, including regression, tree-based models, and neural networks, and applied that knowledge to solve a genre classification capstone project using LSTM and GloVe embeddings PROJECTS

• Developed a book recommendation system using Goodreads data as a course project at Georgia Tech, leveraging Matrix Factorization to provide personalized recommendations based on user preferences Sep 2024 - Nov 2024

Contact this candidate