I'm a graduate from the MS in Analytics program at Georgia Institute of Technology specializing in computational data science with a GPA of 3.91. I earned my Bachelor's degree from USC Marshall in May 2022 with honors and a GPA of 3.74. I standout with my knack for understanding the business landscape, allowing me to effectively contextualize and address various problems. This ability is further enriched by my expertise in SQL, Python, R, and Tableau, empowering me in data manipulation, visualization, and machine learning applications. I am actively pursuing exciting data scientist and data analytics opportunities starting 2024.
Financial reports offer critical insights into a company's operations, but require costly manual review. To address this, we leveraged finetuned Large Language Models (LLMs) to distill key indicators and operational metrics from these reports basis questions from the user. We leverage the FinQA dataset to fine-tune both Llama-2 7B and T5 models for customized question answering. We achieved results comparable to baseline on the final numerical answer, a competitive accuracy in numerical reasoning and calculation.
Details GitHub
We apply a physics-informed deep-learning approach the PINN approach to the Black-Scholes equation for pricing American and European options. We test our approach on both simulated as well as real market data, compare it to analytical/numerical benchmarks. Our model is able to accurately capture the price behaviour on simulation data, while also exhibiting reasonable performance for market data. We also experiment with the architecture and learning process of our PINN model to provide more understanding of convergence and stability issues that impact performance.
Details GitHub
Robo Chef is a program that takes a picture of food, recognizes its recipe, and gives recommendations based on the photo and previous user interactions. We first used a ResNet50 CNN architecture with transfer learning for image recognition. Then we built a food recommendation system based on SVD algorithm and NMF(non-negative matrix Factorization).
Details GitHub
The project involved the development of a predictive model for accurate detection of Brent futures price structures. This was achieved by utilizing Principal Component Analysis (PCA) and a Hidden Markov Gaussian Mixture model. The aim was to capture and analyze market behavior to make more informed trading decisions.
Details GitHub
To predict the 7-day sales forecast for slow-moving items, an ensemble framework was constructed. This ensemble framework combined the power of three distinct models: Facebook Prophet, MSTL decomposition, and Hidden Markov models. The integration of these models led to the achievement of an impressively low Mean Squared Error (MSE) of 2.21. Importantly, this MSE was 58.23% better than that of the Null model, demonstrating the effectiveness of the ensemble approach in improving sales forecast accuracy and highlighting its potential for enhancing decision-making in inventory management and sales planning.
Details GitHub
A robust forecasting model was developed in this project, utilizing Gaussian Process regression, a Bayesian machine learning framework, and fine-tuning the model using PyTorch. The model underwent experimentation with various kernels, including the RBFKernel, during training. This approach allowed for the creation of a highly adaptable and accurate forecasting tool capable of capturing intricate data patterns and relationships.
Details GitHub
This data analytics and visualization project aims to determine the most critical contributors to a country's happiness score. Based on the world happiness report dataset from 2015 to 2022. I conducted regression analysis, feature engineering, and Tableau visualization to show my results.
Details GitHub