projects

Battery and Energy Storage Project Analysis Tool

Streamlit, Python, Techno-Economic Analysis | MIT Energy & Climate Hackathon

• Developed a Streamlit-based techno-economic analysis dashboard to evaluate battery storage projects using inputs such as industrial load curves, RTO market prices, battery capacity, degradation rates, round-trip efficiency, capital and O&M costs and project lifetime
• Implemented modeling to estimate lifecycle costs, savings, and payback period under different operational scenarios, enabling users to compare battery sizing and market participation strategies

Data Center Carbon-Aware Scheduling with Co-Located Renewables

System Design, PJM Grid, BESS, Solar, Wind

• System design for a 100 MW flexible load data center using solar, wind and BESS to shift compute loads
• Designed based on PJM's CO2 emissions modeled to reduce emissions by 33% or 100K ton CO2e annually

Invasive Species Monitoring

Convolutional Neural Networks, Tensorflow, Keras, numpy, pandas

• Explored the biological consequences of globalization and the challenges posed by invasive species for a term paper
• Applied a CNN model to a Kaggle Dataset, achieving a 75% accuracy in identifying hydrangeas

NIR Spectroscopy based Apple Classification

Machine Learning, sklearn

• Successfully reproduced the findings of a paper predicting apple origin through spectral data from NIR spectroscopy
• Reduced dimensionality via Principal Component Analysis, and trained a Support Vector Machine model achieving 100% accuracy in binary classification of apples

Nand to Tetris

Build a Modern Computer from First Principles

• Built components like logic gates, adders, multiplexers and demultiplexers, Arithmetic Logic Unit (ALU) and a CPU

Simple Markov Chain

Takes in a text file as input to create a simple markov chain. Based on the input corpus, generates a sequence of words.