Interests
In my free time, I like to build and explore things that make my life a little easier or more meaningful. I enjoy creative projects, making music, and gardening.
US Economy Pulse PipelineGitHub ↗
An end-to-end data engineering project spanning every layer of a modern data stack—from API ingestion and database storage, to dbt transformation layers, statistical anomaly detection, CI/CD, data quality checks, and a live analytics dashboard.
I built it entirely on my own, tracking six US economic indicators sourced from the Federal Reserve Economic Data (FRED) API. Every layer is something I designed and implemented myself. The project is fully open source.
Tooling
- Python with rate-limited FRED API client
- dbt for transformation (raw → staging → intermediate → analytics)
- Supabase (PostgreSQL) with idempotent upserts
- GitHub Actions for daily orchestration and Slack alerting
- Streamlit Cloud dashboard with KPI cards and recession-risk timelines
- Hybrid data quality: dbt tests + Python anomaly detection
EasyPeasy Cardio
EasyPeasy Cardio is an IOS app as a personal experiment. I wanted something that made it easier to move on days when motivation is low.
I built it around the idea that showing up matters more than intensity. The app keeps things simple on purpose—short sessions, minimal choices, and no pressure to do more than you can.
Features
- Configurable, approachable micro cardio sessions
- YouTube API integration
- Time-based completion instead of performance tracking
- Gentle prompts without guilt or pressure
- Minimal UI to reduce decision fatigue
Tooling
- iOS / SwiftUI
- Simple timer and state logic
- Local-first design
Lyric Video GeneratorGitHub ↗
Lyric Video Generator is a personal tool I built to make releasing music less fragmented. I wanted a single place to sync lyrics to a song and turn that timing directly into a finished lyric video without bouncing between multiple apps.
I built it around the idea that lyric timing is the core work. Once the timing is right, everything else—visuals, animation, export—should follow naturally. The app keeps the workflow linear and focused: write lyrics, sync timing, choose a style, generate the video. It's fully open source—free for any musician or creator to use.
Features
- Musixmatch-style lyric timing using the space bar
- Fine-grained timing adjustments per line (±0.1s / ±0.5s)
- Live audio playback and per-line preview during sync
- Visual templates with configurable fonts, animations, and backgrounds
- Floating emojis and album art thumbnail support
- Automatic intro card for song title and artist
- Built-in YouTube-style outro (like / subscribe / comment)
- Exports tuned for YouTube, Shorts, Instagram, and TikTok
- Optional thumbnail generation during export
Tooling
- Web app (React / TypeScript)
- Audio waveform with real-time playback
- Custom timing and rendering pipeline
- FFmpeg-based video generation
- Local-first project workflow
AI Cooking Companion(WIP)
I’m building an AI Cooking Companion for the way people actually cook—messy notes, half-remembered ideas, substitutions, and hands-full moments. Most recipe and meal-planning apps feel rigid and high-friction: too many forms, too much clutter, and very little personalization.
This project explores a more effortless lifecycle: idea → recipe → meal plan → shopping list → execution—supported by an AI assistant that can work hands-free and learn preferences over time. The vision is a kitchen companion that feels simple, adaptive, and genuinely helpful.
Features
- Zero-friction input (voice + gesture)
- Personalization through learned preferences
- Real-time guidance while cooking
- Meal planning and shopping list generation
Tooling
- LLM APIs for recipe interpretation and generation
- Speech-to-text for voice-first input
- Prompt and interaction design for freeform, messy inputs
- Lightweight backend services for personalization and state
- Screen recording and gesture-based controls for hands-free navigation
- Early exploration of meal planning and shopping pipelines