Dhyey's Projects

Personal Projects

LogFlow AI: Intelligent Logs Analyzer to Drive a Proactive Approach to Businesses At-Large Scale

• Won 3rd Prize at HackMIT Pitch Challenge backed by Y Combinator with guaranteed 1:1 Office Hours

• Coded the frontend and backend of the system MVP to drive insights and automated notifications from the service logs data

• Integrated OpenAI API for seamless functioning of the product, and constructed design docs to allow for faster development process

GitHub Repository

MinervAI: Educational Tooling for Educators for Simulated AI Agents Based Insights into Diversity, Equity, and Inclusion

• Won Grand Overall Prize worth over $3200 at HackUMass XI Hackathon including a $375 worth special letter from Wolfram Alpha

• Developed backend agents' OOP framework using Semantic Kernels from Microsoft

• Drew insights from my experience in successful projects like LogFlowAI and MammothEdu for enhanced development and vision

• Managed GitHub repository, and played an instrumental role in initial system design and conceptualization

GitHub RepositoryDevpost Winner Project Link

ML-based and Pairs Trading on various Stocks and Portfolios

• Implemented ”Buying the Loser”, Decision Tree-based Pairs Trading, LSTM-based Trading and MVP Optimization strategies

• Coded model pipelines in Python while leading the project team of 5 graduate students to nail down an optimal structure

• Calculated and optimized for portfolio risk metrics such as VaR & ES in R and wrote the final report and presented in class

GitHub RepositoryPortfolio Analysis ReportPresentation Pairs Trading

Statistical Trading Strategy in Futures-Markets

• Performed Variance Ratio and Push Response Tests for market behavior on 5-min price data using R and Python

• Implemented and Backtested Trend-following trading strategy with split transactional costs in Python and MATLAB

• Fitted Levy PDF to the distribution of price changes to obtain coefficients and empirically test the fat-tails hypothesis

GitHub RepositoryPresentation

Black Scholes Sentiment To Volatility Trade Recommender

• Engineered a Sentiment based volatility predictor and trade recommender bot using optimization methods in Call/Put Options.

• Used Python, Twitter API and the Black Scholes Model to estimate the fair prices of options based on the top 100 tweets.

GitHub RepositoryReportPresentation

Note Taking App

• Created a note taking web application with edit, delete, and add functionality using HTML, CSS, JavaScript on Netlify

• Scaled to over 100 people; performed quantitative and qualitative UI/UX research by task analysis and writing user stories

GitHub Repository

Calendar Query

• Created a time tracking and data analysis tool for better self-evaluation of productivity, physical and mental well-being.

• Performed Data Analysis, Wrangling and k-Means Clustering using R, Shiny, RStudio, and LaTeX

GitHub Repository

MultiThreaded Java Application for Randomized Account Keeping

• Used distributed computing in Java to make accounts and transactions synchronous and thread-safe

• Used Java, Distributed computing, Threads and Locks, Object Oriented Programming, GitHub and VS Code during the development process.

GitHub Repository

COVID-19 Data Analysis: Shiny and Blog

• Created an app and blog in group of 3 as a part of final project of Data Science class to analyze the periodic and spatial trends of COVID-19, mental health and sentiments in the US.

• Performed Twitter Sentiment Analysis, Spatial Data Superposition and statistical analysis using R, Shiny and Twitter API with 2 teammates.

GitHub Hosted Blog Website GitHub Repository