Software engineering · Data systems · Quantitative research

Harsh Choksi

I build reliable software and data systems for markets, research, and practical decision support.

I am a Computer Science and Data Science student at Rutgers University. My work spans deployed web products, financial data pipelines, machine learning, systematic research, and options analytics.

Email Harsh View GitHub LinkedIn

Portrait of Harsh Choksi — **Rutgers University** New Brunswick, NJ · Class of 2027

3.67 GPA at Rutgers

May 2027 Anticipated graduation

5+ Quantitative research projects

U.S. Citizen Work authorization

Selected work

Products and research with evidence behind them

I take projects from ambiguous questions and raw data through implementation, validation, deployment, and documentation.

Deployed product · Options analytics

Wheel Strategy Options Screener

Live

Designed and deployed an end-to-end web application for cash-secured put and covered-call research using TradingView data and browser-local Robinhood data.

Built Manual and Auto workflows for CSP screening and covered-call positions.
Kept Robinhood credentials and sessions local to Chrome through a custom helper.
Implemented caching, strike selection, assignment rules, sizing, CSV export, and responsive QA.

Delivery: Live web application
Workflows: CSP and covered calls
Privacy: Browser-local credentials

Stack: JavaScript, Node.js, Chrome Extensions, Docker, HTML, CSS

Open live project View source

Realtime AI systems

Voice Agent & Evaluation Harness

Built a Python voice-agent system that connects controlled outbound calls to a realtime AI session, streams bidirectional audio, and turns each run into reviewable engineering artifacts.

12: Patient scenarios
Live: Bidirectional audio
Traceable: Transcripts and metadata
Evaluated: Automated findings draft

Includes an explicit destination guard, call recordings, raw event logs, transcript evaluation, and reproducible scenario execution.

Stack: Python, FastAPI, Twilio Media Streams, WebSockets, OpenAI Realtime

View repository

Systematic research

Quant Trading & Research Portfolio

Published market-regime, covered-call, and systematic strategy research using time-aware evaluation and reproducible backtesting.

92.4%: Balanced accuracy on model-generated labels
50+: Backtested trades
>1.2: Backtested Sharpe ratio
~35%: Backtested annualized return

Historical research results are not live performance.

Stack: Python, Pine Script, LSTM, K-Means, Black-Scholes, GBM

View repository

Financial data engineering

Event-Based Earnings Data Pipeline & Modeling System

Built a Python, SQLite, and yfinance pipeline that turns earnings events into structured features and time-valid modeling datasets.

52.4%: Directional baseline
53.4%: Time-validated directional accuracy

Compared logistic regression, random forest, and gradient boosting while reducing look-ahead bias through time-based validation.

Stack: Python, SQLite, Pandas, Scikit-learn, yfinance

View repository

Industry experience

Experience

Information Systems Intern

Global Industrial Company

June 2025 - Aug 2025

Education

Rutgers University

New Brunswick, New Jersey | Anticipated May 2027

Degree Bachelor of Science

Computer Science and Data Science (Computer Science option)

Additional study Minor in Mathematics and Quantitative Economics

Data Science certificate

Academic standing 3.67 / 4.00 GPA

Dean's List

Work authorization U.S. Citizen

Collaboration

Research and engineering teams

Quantitative Finance Club

Rutgers University | September 2023 - Present

Built and backtested systematic strategies in Python, evaluated risk and return, and contributed statistical analysis across more than five quantitative research projects. Selected historical simulations improved by approximately 5%.

Rutgers Formula Racing

Rutgers University | September 2024 - May 2025

Designed an F1-style steering wheel in SolidWorks and analyzed telemetry to support data-driven steering and suspension development under FSAE constraints.

Toolkit

Technical foundation

Languages, tools, and concepts I use to move from raw data to tested systems.

Languages: Python, JavaScript, SQL, R, Pine Script
Data and modeling: Pandas, NumPy, Scikit-learn, SQLite, statistical modeling, feature engineering, time series analysis
Engineering tools: Node.js, Chrome Extensions, Docker, Git, Selenium, Power BI, Excel
Relevant coursework: Data Structures, Algorithms, Computer Architecture, Data Management for Data Science, Machine Learning, Linear Algebra, Probability, Econometrics
Technical concepts: ETL workflows, data pipelines, statistical modeling, time series analysis, feature engineering, risk metrics, options strategies
Certifications: Anthropic AI Fluency: Framework & Foundations, Anthropic Claude 101

Beyond projects

Work history and interests

Volunteering and work

Woodbridge Children's Library Volunteer
Topgolf Food Runner / Caddie

Hobbies

Semi-Professional Table Tennis
YouTube Content Creation

Start a conversation

Interested in software, data, or quantitative work?

I welcome conversations about engineering, analytics, research, and the systems behind Wheel Strategy Screener.

Email Harsh Call Contact form