Section

A cloud-native rApp deployed on the Ericsson Intelligent Controller (EIC) platform. It uses AI/ML-driven predictions to optimize 5G radio network traffic distribution across NR cells — monitoring cell KPIs, profiling network nodes, and managing EN-DC distribution parameters at scale.

Work I did

Worked on scaling the optimizer to 65,000+ cells — distributed architecture, live telemetry ingestion, and an AI prediction model with fallback logic when accuracy degrades. Also built out drift detection and automated recovery to keep the system stable under changing network conditions.

Team I was in

Software developers, ML engineers, testers, and telecom domain experts. Daily scrum, agile workflow, and code reviews through Gerrit.

Organization context

Ericsson is one of the largest telecom equipment vendors in the world, with software running on live 5G deployments globally. My team sits inside the Radio Access Network division focused on intelligent network automation.

Section

Data & IP Research Division — the group that tracks innovation trends across Canada through patent and trade data.

Work I did

The analysts were losing days just getting data ready. Four sources (PATSTAT, WIPO, Statistics Canada, World Bank), all in different formats. I built ETL pipelines that pull everything into one clean store of 2M+ records, which cut their prep time by about 40%.

I also trained NLP classifiers with spaCy and Hugging Face to auto-tag 850K+ patents by technology domain — something that used to be done by hand. And I built a patent search engine combining TF-IDF with embedding retrieval, which improved relevance about 20% over the old keyword-only system.

Team I was in

Economists and policy analysts. A lot of the job was figuring out what they actually needed and building tools that fit how they already work — not just what made sense technically.

Organization context

Canadian Intellectual Property Office (CIPO) under Innovation, Science and Economic Development Canada. CIPO administers patents, trademarks, and other IP in Canada. My work supported the research and operations teams that inform how Canada understands and tracks innovation activity.

Section

Product engineering for Interact Health PRO, supporting a network of rehabilitation centres and treatment facilities across the GTA.

Work I did

Delivered features end to end — REST APIs with role-based access control, real-time clinic availability over WebSockets, and the React frontend. Some routes were slow so I went through the PostgreSQL queries, added missing indexes, and rewrote the worst ones. Got response times from 420ms down to 140ms.

Team I was in

Small team with clinicians in the room. When the people using your software are right there, you think differently about what you're building.

Organization context

The organization coordinates care for pain-related conditions across the GTA, working directly with primary caregivers and clinics. Doctors and therapists collaborate closely to ensure each patient's needs are met — reliability in the software isn't optional.

Section

School of Mathematics & Statistics, undergraduate courses in Calculus and Linear Algebra.

Work I did

Review sessions and office hours over 5 terms for 500+ students. Mostly helping people get past the point where they felt stuck on limits, eigenvalues, and proofs — the stuff that makes or breaks first-year STEM.

Team I was in

Collaborated with course instructors and fellow TAs to align on teaching goals and provide consistent support across sections.

Organization context

Carleton University is a leading Canadian research university. These foundational math courses serve students across every STEM program — getting them right matters.