About

Jivitesh Poojary is a skilled Machine Learning Engineer with a passion for working on complex data analysis processes and developing intuition in data modeling. He holds an MS in Data Science (Honors) and a BTech in Information Technology. Jivitesh has a deep understanding of machine learning algorithms and their applications in different domains. He enjoys working on the whole spectrum of a data science project, from understanding the business needs to presenting results to decision makers. With experience in various roles as a Data Scientist, Machine Learning Engineer, Data Engineer, and Data Analyst, Jivitesh can anticipate challenges and estimate effort required for a successful data science implementation. He has a wide range of technical skills, including programming languages such as Python, R, Scala, Java, and SQL, and machine learning tools such as Scikit-learn, Jupyter / IPython, RStudio, NumPy, Pandas, SciPy, and Tensorflow. Jivitesh is also proficient in data engineering tools such as Spark, Hadoop, Hive, Pig, Oracle PL/SQL, PostgreSQL, IBM DB2, and MongoDB, and BI and visualization tools such as Tableau, Shiny, Matplotlib, GGplot, D3.js, Seaborn, Bokeh, Vega, and Excel. He has experience working with natural language processing tools such as Lucene, Stanford CoreNLP, Gensim, Spacy, and NLTK. Currently, Jivitesh works as a Senior Machine Learning Engineer at Dish Network, where he primarily works with Customer Retention and Dish Media teams to provide data science solutions that help business units be more efficient and drive growth. He has exposure to industry-standard data sets provided by demographic data aggregators and credit rating agencies. Jivitesh has developed a model to predict the propensity of customers to churn within the next business cycle, reducing the feature space to 1/5th of the original size and optimizing the ML pipeline for scaling, rigorous testing, and feedback loop. He also created an application to generate viewership forecasts on a Quarterly and Weekly level using Pandas UDF and Facebook Prophet, which enabled the business to obtain 80% better forecasts on average and freed up time of 3 months/year/analyst. Additionally, Jivitesh designed and deployed an impression capacity forecast tool on AWS, developed a reporting procedure to track the efficacy of third-party ad targeting platforms, and developed CI/CD pipelines with Gitlab,