Ankit Beladiya

ANKIT BELADIYA DATA SCIENTIST : +1 (514)- :- :Montreal, QC :ankit-beladiya SUMMARY • • • • • • Proven industry experience in building and deploying AI driven projects from inception to production Solid understanding of Machine Learning, Deep Learning and Statistics Advanced programming skill in Python, SQL and Scala Proficient in using Big data technology such as Spark, Hive and MongoDB Excellent communication and interpersonal skills gained through professional work experience Problem-solver, Organized, Detail-oriented, Autonomous, Self-starter, Quick learner TECHNICAL SKILLS Machine Learning Python Spark DB DevOps AWS Classification, Regression, Clustering, Feature Engineering, Hyper-parameter Optimization NumPy, SciPy, Keras, Tensorflow, Pandas, Matplotlib, XGboost, Scikit-Learn, Dask, seaborn PySpark, MLlib, SparkSQL PostgreSQL, MongoDB, Hive Docker, Docker Compose, Kubernetes S3, EC2 EXPERIENCE Stradigi AI Jan 2019 – Mar 2020 Data Scientist 1. Automated Stock Trading Platform • Build an automated trading platform using combination of unsupervised, supervised and deep learning ML models • Achieved higher returns than SPX index with better Sharp and Sortino ratio • Problem formalization based on client’s requirements • Performed data collection, cleaning and storing in PostgreSQL database • Data exploration, transformation, visualization and reporting • Data parallelization using Dask framework • Wrote a detailed report to client communicating our findings and results • Build and deploy models on AWS EC2 using Docker • Collaborated with the solutions team to productionize the software 2. Image Segmentation Project for Health Care Client • Built an image segmentation pipeline using deep learning model for a health care client • Data augmentation, model training and serving using Tensorflow 3. Kepler Platform - Data Science Pipelines • Used Spark for structured and unstructured data ingestion • Built production-grade data science pipelines for the time series forecasting, NLU and image processing • Hyper parameterization of classical machine learning models and neural networks 4. Teamwork and communication responsibilities • Provided training to other staff on machine learning and big data concepts • Followed software development best practices and an agile methodology EDUCATION Master of Engineering (Electrical) University of Windsor - Ontario, Canada May 2017 – Sep 2018 ANKIT BELADIYA| 1 PROJECTS Sentiment Analysis with LSTM • Stored live tweets in PostgreSQL database • Tokenized html tags, special characters and emojis using Regular expression • Converted tokens into embeddings using glove embeddings • Built and trained a LSTM network in TensorFlow • Monitored training of model using TensorBoard • Evaluated model performance using the Confusion matrix • Wrote clean and reusable code in python Recommendation System with Big Data • Ingested user data into Hadoop file system (HDFS) • Read the data from the HDFS using Apache Spark • Data modelling and pre-processing using PySpark and Spark SQL library • Applied the Collaborative Filtering Technique to recommend new products using Spark MLlib • Stored results in Hive Database Jul 2018 Apr 2018 Missing Data Imputation Aug 2017 • Implemented fuzzy clustering-based algorithm using Scikit Learn • Used Elbow method to find number of clusters in data • Imputed missing data using Expectation Maximization (EM) method • Compared algorithm’s accuracy and performance against K-means and Gaussian Mixture Model (GMM) clustering • Implemented python API for easy use of the algorithm CERTIFICATES Coursera Applied Data Science with Python Sep 2018 Coursera Machine Learning with Big Data May 2012 ANKIT BELADIYA| 2