CHIRANJEEVI VEGI

ac4k1u@r.postjobfree.com www.chiranjeevivegi.com +1-734-***-**** Ann Arbor chiranjeevivegi cjvegi SUMMARY

Focused on drawing insights from messy data, building machine learning models and data storytelling. My background is in data wrangling, data analysis, machine learning, deep learning, data-driven modeling, predictive modeling, supply chain design, and analytics, and smart manufacturing

EDUCATION

Springboard

Data Science 2018

Mastering the data science process, from statistics and data wrangling, to advanced topics like machine learning and data storytelling, by working on real projects

University of Michigan, Ann Arbor

Masters Mechanical 2017

Emphasis on Data science in Manufacturing

EMPLOYMENT

University of Michigan and Ford Motor Company, Graduate Student Researcher, Livonia, US Feb 2017 Dec 2017 Project Poster. It involved following skill set

1. Data wrangling. Data size: 8.98 GB

2. Inferential statistics (Hypothesis testing, Confidence Intervals etc.,) 3. Finding hidden patterns in the data

4. Data storytelling and visualizations.

PepsiCo, Assistant Production Manager, Bengaluru, India Jun 2014 Mar 2016 While working in PepsiCo, I led a team of 300 people. During this phase using historical data, 1. Evaluated and improved manufacturing methods which increased net efficiency by 5.2% 2. Investigated equipment failures to diagnose faulty operation and made appropriate maintenance recommendations which resulted in downtime reduction of 13%

SCRIPTING LANGUAGE: Python, R, HTML &CSS (some exposure) MACHINE LEARNING: scikit-learn, TensorFlow, Keras, Tensorboard, NLP VISUALIZATION: Tableau, matplotlib, seaborn, ggplot DATA: SQL, Webscraping

SKILLS

PROJECTS

Statistics Online Computational Resource (SOCR) Big Data & Predictive analytics Dec 2017 Current Objective: Approaching biomedical and health science research from the perspective of Big Data applications in nursing informatics, multimodal biomedical image analysis

Skills: Higher dimensional visualization using TensorBoard, TensorFlow, Unsupervised Learning Toxic Comment Classi ication Challenge Jan 2018 Current Objective: To build a multi-headed model that’s capable of detecting different types of toxicity like threats, obscenity, insults, and identity-based hate better than Perspective’s (https://perspectiveapi.com current models (Kaggle Competition) Skills: web deployment, NLP, Machine Learning, Multi-class labels Porto Seguro's Safe Driver Prediction Oct 2017 Nov 2017 Objective: Using data, to predict if a driver will file an insurance claim next year (Kaggle Competition) The challenging aspect of the project is dealing with imbalanced data. Finished in top 17% on the kaggle Leaderboard Skills: Data Wrangling, Exploratory data analysis, Inferential Statistics, Machine learning (Xgboost, SVM, Adaboost, Random Forest) and Hyperparameter tuning

BLOGS AND CERTIFICATES

Medium Blog

Data Science Certificates