lucas.pirvuti.com
Work About Contact Resume
SWE Intern @ CanDeal · Toronto

Lucas Pirvuti

CS at Waterloo, Business at Laurier. Interested in quant trading and market structure.

View work → GitHub LinkedIn
geometric brownian motion · live
σ {{ sigmaLabel }}   μ {{ muLabel }}   paths {{ pathLabel }}
move your cursor to change volatility

Selected work

03
Jul 2026 — present Quant research

Signal Desk

A chart-pattern scanner, backtester, and trade journal covering 92 tickers across five timeframes. I benchmarked the detectors against matched random-entry controls, found no edge, and said so in the README — the honest result became the point of the project.

11
Pattern detectors
92
Tickers scanned
160
Unit tests
  • Caught a backtest bug that scored trades which never opened, then retracted the conclusions it broke
  • Pooled thin per-symbol samples with empirical Bayes shrinkage, tuned on out-of-sample data
  • Strict-prefix windows so detectors never see the future; pessimistic fills throughout
  • Rate-limited market feed kept alive with request caching, retry backoff, and stale-value fallback
TypeScript Next.js 16 React 19 Tailwind v4 Yahoo Finance API
View repo
Aug 2026 Numerical methods

Green Simulation Limits

Reproduced a 2022 Monte Carlo method that reuses samples across scenarios and confirmed its 100× error reduction — then found the advantage vanishes by eight dimensions once the density it depends on has to be estimated from data.

100×
Error reduction
8d
Breakdown point
2M
Validation draws
  • Error measured against a closed-form benchmark, itself checked on 2M brute-force Monte Carlo draws
  • Control estimators blind to the conditioning features confirmed dimensionality caused the loss
  • Negative result written up as the finding rather than tuned away
Python NumPy SciPy pytest
View repo
Dec 2025 — Jan 2026 · Search & rescue

Waypoint

Built at DeltaHacks 2026 with a small team. A search-and-rescue simulator modelling 1000+ lost-hiker agents across five ISRID behaviour strategies over NASA SRTM terrain; I built the Monte Carlo engine, parallelised it with ProcessPoolExecutor, and served slope-adjusted heatmaps through async FastAPI.

1000+
Agents simulated
32
Parallel timesteps
Python FastAPI NumPy Rasterio Mapbox GL JS
Team repo

Experience & education

May — Aug 2026

Software Engineer Intern

CanDeal · Toronto, ON
  • Refactored the fixed income security master configurator and datastore for richer analytics, extending coverage across 7 instrument types — callables, amortizers, FRNs
  • Integrated a QuantLib-based C++ library for price, yield, and PV01/DV01 conversions
  • Rebuilt real-time security master synchronisation between source and target environments
  • Built Claude agent skills that configure complex FIX-based trade negotiation message flows
  • Rebuilt the observability layer on Grafana/Prometheus with custom TibcoRV event sources, monitoring pricing flows spanning 10M+ messages per day
Education
University of Waterloo
BCS, Computer Science
Sep 2025 — Apr 2030
Wilfrid Laurier University
BBA, Business Administration
Sep 2025 — Apr 2030

Double degree — CS and business, concurrently. Expected graduation 2030. Cumulative average 81%.

Euclid 2025 Honour Roll — top 2% of 27,000
President's Scholarship of Distinction
Qualified for the President's Research Award
Stack
languages
Java 21 · Python · TypeScript · C++ · SQL (MySQL, Postgres)
frameworks & libraries
Spring Boot · QuantLib · Vaadin · Next.js · React · FastAPI · Tailwind · NumPy · SciPy · Pandas
tools
Git · Maven · Grafana · Prometheus · TibcoRV · Linux/Bash · Vim · pytest
concepts
FIX Protocol · Observability · Concurrency · Monte Carlo simulation

Looking for Summer 2027 SWE internships.

Happy to talk about simulation, quant research, or fixed-income market structure. Fastest way to reach me is email.

lpirvuti@uwaterloo.ca LinkedIn GitHub
© 2026 Lucas Pirvuti Toronto, ON