Software Engineering

Designing a C++ foundation for simulating rocket flight dynamics using custom 3D vector mathematics, including position, gravity, forces, and navigation-related calculations.

Hack the Future 2026 build: bridges daily to-do lists and weekly goals via Google Calendar/Tasks, deterministic urgency scoring, Groq JSON prompts for importance and coaching, Eisenhower Matrix UI, and calendar slot suggestions with local fallbacks (see repo docs/llm-integration.md).

Public integration layer I built from the upstream SpenderQ framework (galactic-ai/SpenderQ) for the UW Bothell quasar research group: a data-driven Python stack that embeds the Spender convolutional spectrum autoencoder to reconstruct intrinsic quasar continua, mask absorption (including the Lyman-α forest), and batch-process large spectroscopic datasets. Complements the group’s EHVO analysis pipeline as an ML-based continuum fallback when classical normalization is insufficient.

Built overnight at TechTogether Seattle: a mobile app that slices the day into 'time blocks' and uses machine learning to learn routine patterns, suggest Pomodoro-style breaks, and auto-reschedule the rest of the day when a task runs long. Won the 'Planting the Seed for Growth' award.

In rural France I designed a custom Excel-based crop management system for a working farm's greenhouse: a color-coded grid modeling every planting bed by crop category (leaf, root, fruit, aromatic) and status (seeded, planted, completed) to optimize planting schedules, yield tracking, and resource allocation across dozens of concurrent rotations.

Prototyped a way to infer whether a scanned Wi-Fi network corresponds to home, work, or school by clustering visit frequency and time-of-day patterns — aimed at reducing noisy location pings without extra user input.

Lead backend developer for a T-Mobile DevEdge hackathon build: a geofenced anti-theft system that flags a bike as stolen the instant it leaves a locked zone, streams live location to a companion iOS app, and fires an automated SMS alert to the rider.

React Native front-end for controlling mock smart-home devices (locks, bulbs) over a shared IoT API, with animated on/off interactions and live device state fetched on mount.

A diagnostic web tool that decodes raw hex-string UDP payloads from field devices into readable telemetry (temperature, battery level, speed, status) and outputs live service health, cutting down the team's manual testing time.