CS @ University of Waterloo '28
I build things that help real people. Starting with a laptop was the only option — turns out that's enough.
open to: cloud · infra · devops · swe · mlops — summer & fall 2027
| Role | Company | Focus |
|---|---|---|
| ☁️ Cloud Engineer Intern | Definity Financial | AWS cost optimization, FinOps, infrastructure |
| 🤖 Software Engineer, AI | Hands AI | RAG pipelines, LLM systems, monetization |
| 🛠️ Software Engineer, Infra | WATonomous | HPC clusters, Kubernetes, CI/CD, observability |
| WARG | 3D localization, drone systems |
- 70% AWS compute costs at Definity → 118 idle hours/week eliminated
+ 40% RAG retrieval accuracy → 7,000+ items, hybrid pgvector + BM25
- 80% non-compliant recommendations → GPT-4o-mini post-retrieval filtering
+ 10x vector search precision → FlashRank re-ranking + HNSW indexing
! 25+ daily active users → Goosedyssey Chrome extensionGoosedyssey — Chrome extension that transforms UWaterloo's exam schedule into something actually readable. Color-coded midterms, finals, reminders. 25+ students use it daily. Some offered to donate. That's the one I'm most proud of.
Snowspace — Geospatial analysis tool using Google Earth Engine + Sentinel-2 satellite data. Custom KD-tree over 30k+ coordinates → O(log n) lookup, ~28k ops/sec. Top 5 / 200+ teams at BramHacks. Attracted VC interest.
SyncSpace — Collaborative whiteboard for 30+ concurrent users. Y.js CRDTs over WebSockets, conflict-free real-time sync, 40% memory reduction via Ramer-Douglas-Peucker path simplification.
NeuroNotes — Multi-stage RAG pipeline with FlashRank re-ranking and HNSW indexing. 10x retrieval precision, <100ms vector search. Local Llama 3.2 via Ollama.
Languages
Python TypeScript JavaScript Java SQL C++ Bash
Cloud & DevOps
AWS GCP Docker Kubernetes GitHub Actions Prometheus Grafana Linux
Frameworks & Data
React SvelteKit FastAPI Node.js PostgreSQL Supabase Redis GraphQL
AI & Systems
RAG/LLM Pipelines pgvector ChromaDB Y.js CRDTs Distributed Systems Microservices