I build apps, AI tools, and research software. My work spans native mobile development, local AI, browser-agent tooling, and reproducible machine-learning experiments.
Based in Winnipeg, Canada. I also lead an AI club at St. John's-Ravenscourt, teaching students and teachers to turn ideas into working software.
R茅sum茅 路 Project catalog 路 Writing 路 Email
| Project | What I built | Explore |
|---|---|---|
| Evidence Board | A research workspace with source-linked claims, ten browser-agent tools, reviewable proposals, and undo. The public sample uses illustrative data. | Code 路 Demo 路 Tool contracts |
| Context | An offline relationship-memory app with reviewed imports, optional on-device AI, and native iOS bridges. | Code 路 Screenshots 路 Privacy |
| BlockScore | Founder and product lead of a five-person team building a SwiftUI app for Canadian address research. Reports show sources and coverage; the app is in development and source is private. | Public data companion |
| Project Hope | A charity operations platform with a desktop sample, contact migration, organization permissions, and local AI assistance. Real team use requires a hosted workspace. | Code 路 Website and downloads |
| WildGap | A biodiversity survey tool using H3, GBIF, and weather data to identify observation gaps and plan field missions. | Code 路 Demo |
| Clearcoat Quote Studio | A vehicle-photo quoting MVP with browser quality checks, deterministic pricing, and human-review routing. | Code and screenshots |
- AgriStress: co-founder and lead developer of a multispectral stress-mapping research platform. Working with University of Manitoba Plant Sciences researchers to incorporate longitudinal wheat and canola drone imagery. Research prototype; code is private.
- Quant Systems Lab: chronological order-book replay and independently checked accounting. The published study processes 301,587 messages; all five policies lose money on the sample. Study.
- UFC Predictor: pre-fight features and probability evaluation. The recorded chronological benchmark reports 63.55% accuracy on 1,709 held-out fights. This is a historical result, not live betting performance. Results.
More work includes municipal-data tools, accessibility testing, adaptive-model research, engineering simulation, and business software. See the full catalog for status and evidence.
I use AI coding tools throughout development. I set the product direction, review changes, and test the result. I care about source provenance, useful interfaces, and making failure cases visible.
Languages: Python, TypeScript, Swift, Java
Tools: React, React Native, Next.js, Django, FastAPI, NumPy, scikit-learn, PyTorch, Playwright, GitHub Actions
Honours student at St. John's-Ravenscourt, graduating in June 2027. My experience includes product strategy, research, teaching, and running a commission-based resale business.
IBM SkillsBuild Artificial Intelligence Fundamentals certification. Second place in an RRC Polytech business competition.
Last reviewed: 12 October 2026. Project links distinguish working samples, research, and unfinished releases.