Mantle
Worked on the Go pipeline that turns unstructured financial PDFs into a searchable vector database, and the chat interface people query it with in plain English.
I'm in my last year of Waterloo Engineering. Most of my work has been AI and data pipelines, plus the interfaces that sit on top of them.
Worked on the Go pipeline that turns unstructured financial PDFs into a searchable vector database, and the chat interface people query it with in plain English.
Performance attribution across billions in AUM. Replaced fragile spreadsheet work with pipelines and built the dashboards portfolio managers now use.
Built a dashboard that connects car dealerships to prospective buyers through generated leads. Moved reads onto AWS replicas once search started to suffer at peak traffic.
My first co-op: halved Docker image build times with multi-stage builds and layer caching, and automated the deploy pipeline with GitLab and Bash.
Worked in the Health AI & Analytics Lab on how clinical delays affect outcomes for stroke patients.
Led the web software side of Waterloo's hyperloop team: schema design, API architecture, code review, and mentoring.
My first algorithms course: searching, sorting, stacks, queues, trees, and graphs. Introduced space and time complexity analysis, which has come up in every interview since.
Mostly raw SQL, plus UML modeling and database normalization for real-life scenarios. Also an introduction to full-stack work with React and Node.js.
A team project building a full-stack application with React, Node.js, and MySQL, run with agile methodologies and test-driven development (TDD).
Second algorithms course: greedy, dynamic programming, and graph algorithms. Favourite part was the NP-hard and unsolvable problems, and the approximations and heuristics you use on them anyway.
Designing interactive systems start to finish: user requirements analysis, information and interaction design, prototyping, and evaluation. Prototyped in Figma.
Supervised and unsupervised learning, with the emphasis on training and testing models honestly. Course project: Built and compared different models in Python to predict heart disease.
Text retrieval and web search: retrieval models, index construction, and evaluating result quality with t-tests. Course project: Built a search engine in Python using BM25.
Human information processing: attention, memory, pattern recognition, language, decision making, and problem solving. Explained more of what people do day to day than anything else I took, and I kept spotting the concepts outside class.
Integration techniques, differential equations, parametric and polar curves, infinite series, and Taylor polynomials. The first time math was fun rather than following a recipe that always works.
Badminton most weeks and skiing whenever there's snow. After terms, I like to travel to recover from sleep deprivation and/or exam hell ✈️

The backstreets were incredibly relaxing to walk through 😌

Karaage, chashu and a soft-boiled egg in a spicy broth. Side of pan-fried gyoza. Best meal I've ever had 😋

A winter must-do: skiing on the weekends 🎿

Escaping the Canadian rain for a week of sun and sand 🏖️