Ethan Xia

Relevant Coursework

• CIS 1200: Programing Languages and Techniques
• CIS 1210: Data Structures and Algorithms
• CIS 1600: Discrete Mathematics
• CIS 2400: Introduction to Computer Systems
• CIS 2620: Automata, Computability, and Complexity
• CIS 4190: Applied Machine Learning
• CIS 4210: Artificial Intelligence
• CIS 5450: Big Data Analytics
• CIS 5710: Computer Organization and Design
• EAS 2030: Engineering Ethics
• EAS 5450: Engineering Entrepreneurship
• NETS 2130: Crowdsourcing and Human Computation
• MATH 1400: Single Variable Calculus
• MATH 1410: Multivariable and Vector Calculus
• MATH 2400: Linear Algebra and Differential Equations
• STAT 4300: Probability and Statistics
• ECON 0100: Introducory Economics: Micro
• ECON 0200: Introducory Economics: Macro
• ACCT 1010: Accounting and Financial Reporting
• OIDD 2550: Artificial Intelligence, Business, and Society

Work Experience

• AdminifAI
  Software Engineering Intern

  Jan 2026 - Present
  Remote
  • Currently building!
• Bonk
  Co-Founder & CTO

  Nov 2025 - Present
  Philadelphia, PA
  • Co-founded Bonk, a wearable fitness platform combining on-body sensors and cloud-based ML to monitor sweat, heat strain, and metabolic stress, delivering real-time hydration and fueling guidance.
  • Designed and implemented end-to-end data pipelines and AI models for personalized physiological monitoring, using time-series forecasting and baseline modeling to optimize performance metrics.
  • Building and maintaining the full-stack software platform, including cloud-based data storage (Snowflake + AWS), REST APIs, and a user dashboard for real-time tracking and analytics.
• Fundstrat Global Advisors
  Full Stack Developer & Data Science Intern

  Aug 2024 - Aug 2025
  New York, NY
  • Designed and optimized fully automated data pipelines and web scrapers using Python and SQL, processing 4M+ macroeconomic data points daily and eliminating 100% of manual data input while powering insights for 6,000+ clients.
  • Built a scalable analytics system with an AI-driven recommendation engine tailored to research interests and analyzed historical macroeconomic data for investment research, with my insights presented on CNBC.
  • Contributed to the full-stack development of Fundstrat’s production charting platform, working across frontend and backend systems.
  • Reported directly to the Head of Data Science and Research, interpreting macroeconomic data and translating insights into client-facing market reports.
• Greenwich Country Club
  Tennis Coach & Pro Shop Manager

  Jun - Aug 2023; Jun - Aug 2024
  Greenwich, CT
  • Tennis coach for children and teenagers, learning valuable lessons in patience, communication, and leadership.
  • Managed the pro shop, developing strong business, management, and customer service skills.
• Tirrel Corporation
  Full Stack Developer Intern

  Jun - Aug 2023
  New Canaan, CT
  • Worked as a full-stack developer using the Hoon programming language to build a peer-to-peer social media application.
  • Gained hands-on experience with decentralized networking, functional programming paradigms, and the full software development lifecycle, from design and implementation to testing and iteration.

Projects

• ImKomming
  • Built a full-stack, production web application integrating the Wahoo API to analyze cycling workout data, featuring OAuth authentication, serverless APIs, interactive maps, and advanced performance visualizations, mimicking Strav Premium's features.
  • Designed and deployed machine-learning pipelines (LightGBM/XGBoost, K-Means) to estimate Functional Threshold Power and personalize training zones using power, heart rate, and fatigue features.
  • Implemented scalable, serverless data processing and chart generation using Next.js, TypeScript, and Python, enabling real-time analysis of FIT files without persistent storage.
• DraftEdge
  • Built an end-to-end NBA draft prediction pipeline integrating NCAA, NBA Combine, and NBA performance data; engineered conference-adjusted and temporal features to identify undervalued prospects.
  • Developed and evaluated multiple models including logistic regression, XGBoost (with hyperparameter tuning), and player similarity systems (UMAP + k-NN) to predict NBA performance, draft position, and career longevity.
  • Designed scalable data ingestion, cleaning, and feature engineering workflows in Python to process multi-source datasets and produce actionable insights beyond traditional scouting metrics.
• Sync
  • Built a full-stack web application that syncs Spotify playback history with Strava activities using OAuth 2.0, automatically matching songs played during workouts and generating playlists from activity sessions.
  • Developed a Next.js + TypeScript app with serverless API routes, cookie-based authentication, token refresh logic, and robust pagination to handle Spotify and Strava API constraints and playback history limits.
  • Implemented automated playlist creation, Strava activity description updates, and a responsive UI with animated backgrounds; deployed on Vercel with comprehensive error handling and secure user authentication flows.
• Cue
  • Developed an iOS app in Swift/SwiftUI that dynamically curates Spotify music recommendations using real-time heart rate data from Bluetooth LE monitors and Apple Watch (HealthKit), generating the "next best song" based on heart rate zone to match the vibe of the workout and automatically adapting playback to workout intensity zones.
  • Implemented secure keychain access for storing sensitive credentials, integrated Bluetooth LE packages in Swift to establish reliable connections with heart rate monitors, and added background app refresh functionality.
  • Designed a hybrid music recommendation system integrating Spotify Web API playback control with Cyanite AI audio similarity analysis (BPM, energy, mood), backed by a secure Node.js serverless proxy on Vercel to handle GraphQL requests, webhooks, and API key protection.
• Capybara Detector
  • Built a real-time pose detection and classification system using OpenCV and MediaPipe's Holistic model to capture webcam video, extract body/face/hand landmarks, and classify six pose categories with real-time visualization and confidence scoring.
  • Developed an end-to-end machine learning pipeline including an interactive data collection tool, geometric feature extraction from pose landmarks, Random Forest classifier training, and real-time inference for pose recognition.
• ESP32 FPV Drone & Arrow (Currently Building)
  • Currently designing and building a custom FPV drone from sourced components and 3D-printed parts, programming an ESP32 microcontroller for onboard control, wireless communication, and real-time telemetry.
  • Implemented vision-based drone gesture control over Wi-Fi using OpenCV and MediaPipe, enabling low-latency, hands-free navigation through real-time pose and hand-tracking inference.
  • Developing Arrow, an anti-drone trajectory prediction system that discretizes 3D airspace into a matrix-based representation to model and predict flight paths of aerial objects—including drones and ballistic projectiles—using self-developed 3D-CNN and LSTM models.

Personal Interests

Cycling, Tennis, Poetry