CS 440: Intro to AI

Introduction to computer systems that exhibit intelligent behavior, in particular, perceptual and robotic systems. Topics include human-computer interfaces, computer vision, robotics, gameplay, pattern recognition, knowledge representation, planning.

CS 210: Computer Systems

Fundamental concepts of computer systems and systems programming. Hardware fundamentals including digital logic, memory systems, processor design, buses, I/O subsystems, data representations, computer arithmetic, and instruction- set architecture. Software concepts including assembly language programming, operating systems, assemblers, linkers, and systems programming in C.

CS 506: Data Science Tools and Applications

Covers practical skills in working with data and introduces a wide range of techniques that are commonly used in the analysis of data, such as clustering, classification, regression, and network analysis.

CS 541: Probability in Computing

Covers practical skills in machine learning including techniques for clustering, classification, regression, feature selection, and model compression. Emphasizes hands-on application of methods via programming on real- world datasets.

DS 594: Spark! Data Visualization X-Lab Practicum

The Data Visualization X-Lab Practicum offers students an opportunity to learn data visualization skills through course and project-based work. Projects will be completed on a schedule that aligns with topics being covered in class and assignments. This course provides an accurate experience of solving real-world problems with data visualization, and the various tradeoffs that need to be considered. Whether it's how to efficiently use color and space, effectively understand the profile of a dataset or cautiously avoid bias, this course will provide students with a solid understanding of applicable data visualization practices.

CS 330: Introduction to Analysis of Algorithms

Examines the basic principles of algorithm design and analysis; graph algorithms; greedy algorithms; dynamic programming; network flows; polynomial- time reductions; NP-hard and NP-complete problems; approximation algorithms; randomized algorithms.

MA 684: Modern Regression Analysis in R

Application of multivariate data analytic techniques. Multiple regression and correlation, confounding and interaction, variable selection, categorical predictors and outcomes, logistic regression, factor analysis, MANOVA, discriminant analysis, regression with longitudinal data, repeated measures, ANOVA.

CS 320: Concepts of Programming Languages

Concepts involved in the design of programming languages. Bindings, argument transmission, and control structures. Environments: compile-time, load-time, and run-time. Interpreters.

MA 416: Analysis of Variance

Fundamental concepts and analytical skills in analysis of variance, including crossed and nested designs, as well as fixed- and random- effect models. Trend analysis for repeated measures, expected mean squares, and non-parametric techniques. SAS is used throughout the course.

Agentic RAG Chatbot for HPC Support

Engineered a Retrieval-Augmented Generation (RAG) chatbot using ChromaDB and the OpenAI API to support users of Boston University’s high-performance computing cluster. The system integrates scraped documentation and internal files with tool-use capabilities to access real-time system data. Built with FastAPI, the client-server model features SSE streaming, JWT-based session handling, and full interaction logging.

GPU Job Efficiency Analytics

Created a suite of data processing and visualization tools to analyze GPU resource utilization on BU’s HPC cluster. Detected inefficient job usage patterns by analyzing job logs using Pandas and visualizations, helping RCS optimize job scheduling and resource allocation.

Diffusion Embedding Visualizer

Developed an interactive web application to visualize and interpret latent embeddings of diffusion models using K-SAE clustering. The tool helps researchers explore and understand high-dimensional model representations through an intuitive interface.

Gradio Interface for Concept Steering in Diffusion Models

View Gradio Tool

Implemented a Gradio-based web interface for side-by-side comparison of outputs from baseline and modified Stable Diffusion models. Allows researchers to experiment with parameters and visualize concept steering effects in real-time.

CNN Based Reinforcement Learning for Algo Trading

View Presentation

Researching a CNN based reinforcement learning agent trained on candlestick data to trade the market. The underlying models use TensorFlow's sequential and functional APIs. Image generation was optimized by utilizing the numba library in Python. This project is associated with BUFC, and was presented during the end of semester presentations. I am still working on this project, and plan to write a paper to provide detailed information about my research.

Mean Reversion Inspired Algorithmic Trading

View Presentation

Developed and live tested an algorithmic trading strategy for the 2024 Spring semester. Succesfully created a model which significantly beat the market during our live test. Heavily leveraged Jupyter notebooks to develop a strategy that relies on HistGradientBoostingRegressor, a Scikit-Learn model. Presented work to the rest of the club and Alumni Advisory Board.

Live Machine Learning Based Algorithmic Trading

Developed multiple algorithmic trading strategies that leverage machine learning to identify profitable opportunities in the forein exchange markets. Implemented a program hosted on an AWS EC2 instance to trade a variety of assets. The program serves lives predictions on fresh pricing data updated every fifteen minutes. I am still actively working on this project.

Voting Patterns in Boston

View Source Code, Website, and More

Collaborated with a team of student mentees, data scientists, and others at Mass Mutual to analyze voter history and demographics datasets (~10 million rows) to uncover trends among voters in Boston. Our deliverable included an interactive website featuring maps and charts, as well as a concise presentation summarizing the problem and our conclusions to support our client.

Tetris Reinforcement Learning Agent

Learned about various artificial intelligence topics in CS440. The final assignment was to create a reinforcement learning agent to play a modified version of Tetris. Designed features, neural network architecture, reward function, and exploration function. Agent succesfully learned how to play Tetris.

Grid Trading Strategy (Implemented Live)

Created a "grid strategy" that profits off volatility in the foreign exchange market. Due to the nature of the strategy, a large strategy parameter optimization was done. The model was deployed live on an AWS EC2 instance, where it automatically managed orders through the OANDA API. The strategy's performance reached a maximum return of 52%; however, after a large drawdown, I stopped the live trading, ending with a return of 11%.

LLM Powered Notes App

View Github

Made a Tkinter-based note-taking application enhanced with LLM (Large Language Model) capabilities, allowing users to receive autocompletion suggestions based on their inputs. The app allows users to write notes, save files, open existing files, and undo changes.

Algo Trading VIX With Random Forest

View Presentation

Explored various strategies utilizing a RandomForest model to predict the VIX index's next day direction. Backtest revealed promising results for the strategy. Presented work to the rest of the club.