PROFESSIONAL SUMMARY:

• Having 7 years of IT experience as an operations research analyst and data scientist,
• Strong background and experience in data science/ data analysis, machine learning, probability and statistics theory,
• Expert in mining, wrangling and analyzing complex, high - volume and high-dimensional data from varying resources,
• Strong coding skills, performing data cleaning and data transformation activities using Python and R,
• Experience with Big Data ecosystems,
• Extensively involved in data preparation, exploratory data analysis, feature engineering using supervised/unsupervised modeling,
• Solid grasp of Linear/Non-linear regression and classification modeling and predictive algorithms in machine learning,
• Proficient in Regression, Classification and Clustering analysis by using machine learning algorithms,
• Experienced Creating / Modeling Machine Learning systems with Neural Networks and Convolutional Neural Networks with TensorFlow and Keras
• Highly motivated and self-starter with TEMPeffective communication and organizational skills,

TECHNICAL SKILLS:

Programming: Python, R, MySQL, PostgreSQL

Tools: NumPy, Pandas, SciPy, Sci-kitLearn, Pyspark, Statsmodel, TensorFlow, Keras, NLTK, Seaborn, Matplotlib, Plotly, Cufflinks, Choropleth, SQLAlchemny, PySpark, Hadoop, HDFS

Artificial Intelligence: Statistical Machine Learning Algorithms, Linear and Logistic Regression, Decision Tree, K-Means, k-NN, Support Vector Machine (SVM), Random Forest, XGBoost, CNN, OpenCV, YOLO, Faster R-CNN, RNN, LSTM, NLTK, Spacy

PROFESSIONAL EXPERIENCE:

Confidential

• Built a CNN to classify images from scratch by using Keras and Tensorflow,
• Used batch normalization method to develop more efficient model,
• Applied data augmentation techniques to increase training data and create more robust CNN model,
• Used dropout, L2 regularization to overcome over-fitting problem,
• Performed transfer learning, using inception, vgg19, mobilenet and resnet50 model and compared teh model accuracy,
• Developing YOLO, Mask R-CNN algorithms to object detection.

Data Scientist

Confidential, Sunnyvale, CA

• Performed data wrangling, cleaning and preprocessing methods to prepare data for analysis,
• Created new features to build more robust models,
• Analyzed iCloud upload and download performances for photos, drive and backup services for various countries using problem-specific metrics,
• Developed K-Modes, Gaussian Mixture Models (GMM) and Agglomerative Hierarchical Clustering models in order to group countries which has same pattern in regards to iCloud performance,
• Analyzed iCloud customer’s behavior patterns and relationship between iCloud performance and customer engagement and also iCloud performance and revenue.

Confidential

• Analyzed Confidential iCloud data using big data ecosystems such as Hadoop, Pyspark,
• Performed data wrangling, cleaning and preprocessing methods to prepare data for analysis,
• Explored new features by using feature engineering techniques,
• Applied Kruskal-Wallis, Confidential, Chi-square statistical test to detect important features on iCloud Performances,
• Prepared python scripts for iCloud photos, drive and backup services to deploy on Hadoop environment,
• Coordinated with Dev-Ops team to automate statistical analysis daily base,
• Triggered daily emails about iCloud performance in photos, backup, and drive services.

Data Scientist

Confidential, San Jose, CA

• Developed machine learning and deep learning models to provide industrial solutions,
• Organized and conducted machine learning pipeline, with a focus on data preparation, model training and optimization, successfully used visualization techniques to inform and give intuitions to stakeholders,
• Performed exploratory analysis and feature engineering to fit best models on Python,
• Discovered patterns, formulated and tested hypotheses, translated results into strategies which drive growth resulting in increased revenues and improved customer satisfaction,
• Consulting companies and individuals for their data science problems,

Confidential

• Explored teh aspects dat influence teh forecast of client subscription in depositing,
• Feature engineered, handled missing values and created new features by combining existing ones,
• Utilized Python matplotlib and seaborn libraries for visualization and exploratory data analysis,
• Applied Confidential, chi-square statistical tests to determine teh predictive power and association amid teh features,
• Selected teh most significant features to decrease complexity and processing time and increase robustness of machine learning models,
• Created different models including Regression, Random Forest, XGBoost, SVM, and Neural Networks algorithms; compared model performances, and adjusted hyper-parameters to obtain improved outcomes.

Confidential

• Preprocessed data by transforming feature, dealing with missing values, feature engineering, creating dummy variables,
• Explored and visualized data set by using seaborn and matplotlib libraries,
• Developed models with data analysis algorithms such as Logistic Regression, Support Vector Machine, kNN, Naïve Bayes Classifier, Decision Trees, Random Forest and Neural Network,
• Applied Grid Search method for parametric algorithms to find optimum parameters and compared algorithms’performances,
• Ranked top 7% in teh competition.

Confidential

• Analyzed pattern between violence and other features such as gender, ethnicity, religion, age etc...,
• Used categorical statistical techniques, like Chi-Sq test, CramersV, Confidential test, t-test to unveil relation between variables,
• Applied advanced visualization techniques by using RShine,
• Performed spatial analysis on data by using map applications,
• Due to not finding pattern, used permutation test to increase sample size.