Resume

Sign in

Software Engineer Professional Experience

Location:
Dallas, Texas, United States
Posted:
May 30, 2017

Contact this candidate

Abhimanyu Rana

San Francisco, California 469.***.**** ac0j26@r.postjobfree.com

LinkedIn: linkedin.com/in/abhimanyu4 GitHub: github.com/abhi117a Oracle: Java Professional TECHNICAL SKILLS

● Proficient: R, Java, Spring 4.0, J2EE, Python, scikit-learn, NumPy, matplotlib, seaborn, Machine Learning, Natural Language Processing, Hadoop, AWS, MySQL, Oracle SQL, Hive, NoSQL, Git, JavaScript, Node.js, Shell Script

● Exposure: H20, TensorFlow, Theano, Keras, MLlib, Apache Spark, Pig, ElasticSearch, MongoDB, AngularJS PROFESSIONAL EXPERIENCE

Sabre Software Engineer (Machine Learning) Southlake, TX 2016 - 2016

● Engineered GUI customization tools for main flight status platform using Java, Spring 4.0, Swing, and SQL, greatly improving product usability while achieving 20% less reboot cycles.

● Led team of 5 engineers following agile methodologies in developing recommendation system for paid premium services utilizing Machine Learning (collaborative filtering) using Python Scikit-learn, MLlib in Apache Spark. Copart Software Engineer (Machine Learning) Dallas, TX 2016 - 2016

● Developed secure online auction portal for entering markets of Spain, Germany and India using Java, Spring 4.0, AngularJS and RESTful services, expanding company’s business and increasing annual revenue by 13%.

● Improved customer satisfaction by over 90% by engineering recommendation engine to suggest vehicles based on search user patterns using R, and Machine Learning (content based filtering with Jaccard Similarity).

● Solved 25+ user stories and tickets with Java and Spring, improving UX and increasing user retention rate by 8%. Tech Mahindra (AT&T) Software Engineer (Data Science + Applications) 2012 - 2014

● Performed statistical and predictive analysis by applying Machine Learning algorithms using R, Python (Scikit), Holt- Winters, randomForest, achieving insights about valued customers with ~93% accuracy (churn prediction).

● Delivered data visualizations on 2M+ product and service records using R, ggplot, Time Series analysis (ARIMA) which assisted client in devising future market strategies.

● Processed and managed AT&T billing system using Oracle SQL and Java, developing Perl, Shell and PL/SQL scripts to analyze real time transactions and batch processes bringing the application uptime to 99.9%. EDUCATION

M.S. Computer Science (Data Science), University of Texas at Dallas 2017

● Relevant Coursework: Machine Learning (Supervised, Unsupervised, Reinforcement), Neural Nets, Text Mining, Statistical Analysis + Modeling, Data Visualization, Big Data, Cloud Computing B.S. Electrical Engineering, GU 2011

RECENT PROJECTS

Allstate Claims Severity Data Scientist code 2016 Top 15 (out of 500) in Kaggle Competition - Machine learning model designed to automatically predict if an insurance claims can be approved.

● Automated the process of predicting the cost and hence severity of insurance claims with 40+ models like, GLM, GBM, Distributed Random Forest XGBoost, Deep Learning and Ensemble approach on 120 predictors in R (H2O).

● Pre-processed, visualized, cleaned, normalized and hot encoded 600,000+ records of insurance claim data using R. Supervised Learning Data Scientist code 2016

Implementation of various machine learning algorithms to classify and predict if passengers survived the sinking of the Titanic.

● Created statistical analysis on Titanic dataset using R and Python by applying logistic regression and classifiers like Decision Trees, Random Forests, SVM, LDA, QDA, Naïve Bayesian, Artificial Neural Nets and XGBoost in H2O.

● Achieved 94% survivor prediction accuracy by performing Feature Selection (Lasso), measuring accuracy using ROC and Confusion Matrix, and improving models with Cross Validation. Online Store Data Scientist code

Improved UX, increased sales and website traffic

● Implemented Thompson Sampling to find which version of ad is more likely to attract customers in R.

● Improved product location by applying association rule mining (apriori) which increased the sales by 38%. LEADERSHIP + AWARDS

Accolades for Innovation: For Introducing Content Based Filtering to the Copart team. 2016 1st Place, Tech Mahindra Hackathon for developing messaging & geolocating app “QUICK” 2015 Mentor + Team Lead: Tech Mahindra, Sabre 2013 - 2016



Contact this candidate