Pankaj Sharma

As a Data Science Engineer, I believe in the power of telling and finding stories through data. I have 10 years of experience in design and development of machine learning models, data exploration, enrichment, mining, statistical analysis, building data warehouse and visualization of data for Business Intelligence solutions using different machine learning algorithms and Hadoop eco system - HDFS, Hive, Sqoop, Impala, Spark, Python. My current venture is leveraging Google Cloud Platform for deploying ML solution at scale using TensorFlow, Cloud ML engine, BigQuery, BigTable, PubSub, DataProc and Google Cloud Storage.  Developed a regression model in TensorFlow to maximize Oil production by predicting choke valve setting based on input from IOT sensors.  Built analytical pipeline with 2.5 TB data of customer footprints from major retails and Telco on Google Cloud Platform for demographics, behavior analysis and developed different algorithms.  Opioid Care, Ingested data from different sources in cloud storage, data exploration in Datalab and built warehouse in Bigquery and segmented counties.  Developed GBM classification model for DNA mapping of more than 475 labels with 95% accuracy.  Chatbot for Document Search Engine using NLP, Tf-idf and web scrapping for keyword searching using cosine similarity in urls and webpage content.  Applied PCA algorithm to decrease the 2000 dimensions dataset to 120 Dimension with only 0.3% decrease in accuracy.  Implemented deep-learning works with LSTM for sentiment analysis.  Built an algorithm for forecasting of next quarter variance using time series analysis and RNN with an accuracy of 86.2%.  Created framework to build and deploy Jars on cluster built using SBT assembly for different Machine learning modules such as Training, Prediction, Transformation.  Expertise in SK-learn, Tensor flow, Keras and H2O libraries.  Implemented data analytics projects and dashboard development on Tableau and Qlik Sense.