Mihir Sharma

Created Data Architecture for Data Ingestion to provide data to multiple end-points for Analysis & Machine Learning and more. Provided different departments with all the data requirements on a scheduled basis and at different levels. Automated Analysis and report generation on GCP (Google Cloud Platform). Made multiple data pipelines to pull marketing data from sources such as Facebook, Google etc. Provided Insights on marketing that increased revenue by 2.3%.

Conducted data analysis to understand problems. Worked to find the most suitable solutions for issues. Effectively communicate complex information to clients. Utilized algorithmic and programming tools to build helpful predictive models. Created reports and statistical analysis as required.

Mathematically understanding algorithms and their core reasoning to use them for model optimization - implementation using Python. Mathematics involving - Probability and Statistic - analysis using forming sampling distributions, hypothesis testing, maximum likelihood estimation for the best estimate of parameters. Linear Algebra - Handling matrices and tensor for various operations, Gram-Schmidt orthogonalisation, singular value decomposition. Calculus - Linear and Quadratic approximations, convex optimization (Gradient descent, Newton’s method). Algorithms involving - Regression (Linear regression, polynomial regression, Lasso and Ridge), Classification (Naive Bayes, LDA, QDA, RDA, Logistic Regression), SVMs, Decision Trees (Random Forest, Boosting-Adaboost, GBMs), Dimensionality reduction (PCA, LDA), Time Series Forecasting (AR, MA, ARIMA), Natural Language Processing techniques like POS tagging, Frequency Based Vectors.