Hall of Works , Projects
Relics forged in past campaigns.
Lantern of Focus , LUMIO
AI-powered early attention detection and learning support. Tracks student focus in real time and gives evidence-grounded guidance.
A lantern that burns brighter when the mind is present and dims when it wanders. Forged for those who learn differently, so that no student is lost in the fog unseen.
Skill: Vigil , Key feature
Watches attention in real time through the camera, entirely in the browser, and calls upon three counsellors who answer only from trusted scrolls.
Scaling , Skills it draws on
- Computer Vision
- A
- Machine Learning
- A
- RAG
- B
- Automation
- B
Self-assessed: how much this work leans on each discipline, from S (most) down to E.
Forged with , Tech stack
- MediaPipe WASM
- XGBoost
- Ensemble models
- LangChain
- FAISS
- RAG
- n8n
- JavaScript
- Python
Wayfarer's Compass , HIKI
Camping and hiking platform for the outdoors community, with real-time weather, moon phases, location and curated campsites.
A compass that points not north, but toward the next good fire. It reads the sky, the moon and the ground, and remembers every camp worth returning to.
Skill: Read the Sky , Key feature
Combines live weather and moon-phase data with your location to suggest trips shaped by real experience, not guesswork.
Scaling , Skills it draws on
- Web
- A
- Databases
- B
- APIs
- B
Self-assessed: how much this work leans on each discipline, from S (most) down to E.
Forged with , Tech stack
- Symfony
- PHP
- SQL
- REST APIs
- Weather API
- Moon phase API
- Geolocation
Ember Sentinel , Real-time Forest Fire Detection AI
A Random Forest model embedded with Arduino hardware that predicts forest fire risk at the edge from live temperature, humidity and smoke sensors.
A watcher set at the forest's edge, that knows the smell of smoke before the first flame. The keeper of a bonfire, it is said, must also know when fire becomes a threat.
Skill: Early Warning , Key feature
Reads temperature, humidity and smoke in real time over serial and predicts fire risk on the spot, with no cloud needed.
Scaling , Skills it draws on
- Machine Learning
- A
- Embedded
- B
- Data
- B
Self-assessed: how much this work leans on each discipline, from S (most) down to E.
Forged with , Tech stack
- Python
- Scikit-learn
- Random Forest
- Arduino
- PySerial
- C/C++
- Sensors
