01
SCENR
An AI creative agent for your memories: you live the trip, SCENR turns everyone's camera rolls into content that feels like you.
Product manager and project lead for a 10-person student team, and author of SCENR's product and technical specification. Friends upload to a shared trip through an app-less link, and SCENR builds Trip Memory once: the people, moments, places, and best shots across everyone's media. It combines that with each member's style profile, retrieved from their own reference examples, to propose a creative plan the user approves before anything expensive is generated. It then renders a balanced Group Cut and personalized Reels, carousels, and Stories from structured timelines, critiques its own output, and takes revisions in plain language instead of an editing timeline. Late uploads update Trip Memory incrementally and refresh only the outputs they improve, and every generation is costed per person per trip so pricing rests on real unit economics. Planned on Flutter, Node.js, PostgreSQL, Python computer vision, LLM agents with retrieval, and FFmpeg rendering.
- AI Agents
- RAG
- Computer Vision
- Flutter
- Node.js
- PostgreSQL
- FFmpeg
02
NINOS
An always-on cost analyst for cloud and AI spending.
Technical Lead directing 2 engineers (AI and cloud) on a University of Calgary ENGG 503/504 entrepreneurial capstone. NINOS is a read-only AI agent for mid-size companies on Azure or AWS: it sorts cloud and AI spending by team, project, and service, finds waste such as idle servers and unattached storage with a specific fix and estimated saving, and explains in plain English what caused each cost spike. It can look but never change anything, so a person always makes the call. Currently in design: owning technical direction, system design, and stack selection; designing three machine learning models (bill forecasting, spike detection, and a waste detector that learns from accepted and rejected suggestions), each required to beat a simple baseline; and designing the explanation agent behind a replaceable LLM provider layer so no single AI vendor is a dependency.
- AI Agents
- Machine Learning
- Forecasting
- Anomaly Detection
- Cloud Cost (FinOps)
03
FitPic
A fashion accountability platform: daily fits, partner streaks, and a vision model reading your wardrobe back to you.
A computer-vision pipeline on Groq's Llama 4 Scout vision-language model extracts structured outfit attributes from user images against an enforced JSON schema, retrying invalid outputs and storing any that stay unreconciled as null, with 98% schema-valid output and 83% tag agreement over roughly 200 images, backed by PostgreSQL via Supabase Realtime. Behavioral and color trends surface in a personal analytics dashboard, secured behind a serverless Node.js/Vercel endpoint with JWT validation and server-side API-key management.
- React
- Supabase
- Groq API
- Llama 4 Vision
- Node.js
- Vercel
04
NBA Betting
Odds Analyzer
A dual-mode over/under predictor that picks its own best model, every query.
A train-on-demand path trains and evaluates five classifiers (XGBoost, Random Forest and others) per query and selects the best by validation ROC-AUC, alongside a fast path serving the best model logged in MLflow. Model selection is backed by MLflow experiment tracking and a versioned model registry, logging params and ROC-AUC/PR-AUC per run for reproducible comparison across retrains.
- Python
- Scikit-Learn
- XGBoost
- MLflow
- NBA API
