Experience & Projects

Co-founder & Software Lead

Problem: Payers and providers needed transparent, evidence-based predictions for knee replacement outcomes — with explainability and workflow fit.

Approach: Build a production ML pipeline from curated datasets through training, validation, deployment, and iteration driven by real healthcare user feedback.

My contribution: Owned the full ML stack: dataset design and pipelines, calibrated models (e.g. logistic regression with propensity weighting), SHAP explanations, FastAPI + Docker deployment, FHIR/Da Vinci PAS integration, and internal tooling.

Impact: AUC ≈ 0.93 with calibration; SHAP-backed explanations for audits; enabled pilot with a major health insurer; platform ready for InterQual/MCG and ePA workflows.

Electrical Engineering Intern

Problem: Manual testing of relay/driver/thermocouple subsystems was slow, error-prone, and hard to repeat.

Approach: Design and build an automated test fixture with custom hardware and firmware so tests could run end-to-end with logging and a simple operator interface.

My contribution: Designed and built the full fixture: mechanical fixturing, wiring, enclosure, and a custom Altium PCB for communications and power. Wrote embedded C on a microcontroller to automate the test sequence, collect and log data, and provide a user-friendly run/monitor interface.

Impact: Reduced test cycle time by ~4 hours; improved repeatability and reliability of validation.

Read my internship reflection on LinkedIn

Undergraduate Researcher — Electrical System Lead, Autonomous RHIB

Problem: The autonomous hydrofoil RHIB needed a safe, robust electrical architecture and component selection to support autonomous capability and real-world marine operation.

Research: Led extensive market research on motors, microcontrollers, throttles, CAN modules, batteries, and marine-rated components. Read datasheets and technical specs in depth to select parts that met performance, safety, and environmental requirements.

Design: Designed the electrical system architecture for the RHIB with redundancy, electrical safety, and reliability in mind. Ensured the system could support autonomous operation and withstand marine conditions.

Integration: Integrated autonomous capability via CAN communication. Connected sensors, actuators, and compute so the vehicle could operate autonomously while maintaining a safe, maintainable electrical backbone.

Impact: Delivered a production-ready electrical design that enabled CMAR's autonomous RHIB program to move forward with a safe, redundant, and real-world-reliable system.

Watch my CMAR experience video

Undergraduate Teaching Assistant

Problem: Students needed clear explanations and debugging support for embedded concepts and microcontroller code.

Approach: Offer office hours, demos, and structured support; design a final project that tied concepts together.

My contribution: Taught embedded concepts and helped students debug code; developed most of the ECE2564 (Embedded Systems) final project.

Impact: Stronger student outcomes and a reusable, meaningful capstone project for the course.

Undergraduate Researcher

Problem: Mechatronic and robotics systems needed reliable, low-level data from force sensors, encoders, and CAN-based subsystems.

Approach: Design a sensor shield and firmware for force acquisition, filtering/calibration, CAN messaging, and encoder interfaces so control and research code could rely on clean data.

My contribution: Built sensor shield and drivers for force acquisition with filtering and calibration; implemented CAN messaging and encoder interfaces (quadrature and absolute); focused on reliable, real-time data capture for control systems.

Impact: Reliable data pipeline for TREC's mechatronic and robotics projects; cross-disciplinary experience in control, sensors, and electrical subsystem design.