Pranav Subramaniam

large-scale data analysis, databases, data mining, and machine learning techniques (supervised, unsupervised learning, ML models, and optimization). P.hD. in Experimental Physics.

I am interested in analyzing Big Data for the technology and financial industry to use my exploratory data analysis techniques to find actionable insights and solutions from the mass amount of data collected, but not yet interpreted. I have a strong background in multiple programming languages Python, C#, and SQL.

Conduct data analysis using machine learning, and deep learning for hypothesis testing and data exploration and to forecast environmental variables in space and time.

• Developed software for automated cleaning, pre-processing, and training of any new data.

•Tuned hyperparameters to achieve higher accuracy in real-world datasets (test phase).

• Due to its high efficiency and accuracy, this software is still in use in the nuclear physics community.
• Improved setup efficiency by 34% through Monte-Carlo simulations.
• Handled strategic data collection, predicted experimental trends, and reduced costs.
• Prepared impact reports relating to setup performance and overall advancement of new methods.
• Tested a new digital acquisition system using Postgres for analytical applications with SQL queries.

Exploratory Data Analysis using scientific packages (NumPy, Pandas).

• Data cleaning using python libraries (Sci-kitLearn, MatplotLib ).
• Automated data analysis procedure for fitting and checking the goodness of fit (Chi-square, p-test).
• Developed algorithms for error propagation and to get the integral calculus from ROC, and AUC Curves.
• ML techniques - Clustering for identification, regression for fitting the data, and classification for the final analysis of a large dataset.