Yaqzan Ali

Senior Software Engineer · Full-Stack & Cloud

Nine years building and running web platforms: Rails and Python services, React and Vue front ends, AWS and GCP infrastructure managed as code. I lead epics from design to production, keep live systems fast and stable, and mentor other developers.

After hours I build and host my own software: an options trading platform, LLM agent experiments, and a handful of small apps. Most of it is live on this domain, and the source is on GitHub.

[email protected] LinkedIn GitHub London, ON · open to remote (US / Canada)

Experience

Senior Software Engineer · HappyCo

Nov 2022 – present

Multifamily property-management SaaS · remote

  • Own performance and stability for a Rails and React SaaS. I tune slow queries and background jobs, and fix recurring incidents at the root with tests that keep them fixed.
  • Run the full AWS stack (EC2, RDS, S3, ACM, IAM, CloudWatch) as code with Terraform and Pulumi. Added a GCP-hosted service and moved CI/CD from CircleCI to GitLab CI.
  • Mentor other developers through code review and design, and brought Claude Code multi-agent workflows to the team.

Software Engineer · Yuhu

2021 – 2022

Acquired by HappyCo, 2022

  • Stabilized the Paysafe payments platform. Found the concurrency bugs behind failed and duplicated charges, fixed them, and reconciled every affected transaction. No double payments since.
  • Led several epics end to end, including a notification system covering email, SMS, and Firebase push.

Software Developer → Intermediate · Info-Tech Research Group

2017 – 2021

IT research & advisory · London, ON

  • Implemented SOC 2 compliance for the myPolicies product and got it through audit.
  • Built Elasticsearch search and Vue.js features for SoftwareReviews, plus LinkedIn OAuth and Okta/SCIM SSO provisioning.
  • Designed a Sidekiq job queue with task prioritization to keep the app responsive under load.

Selected Projects

Bagholders

quant.yaqzan.dev ↗

Options trading platform. Scores about 74,000 symbols nightly with a six-factor technical model, and every algorithm change has to pass a three-stage backtest gate before the live portfolio sees it.

HarnessBench

research

A/B testbed for LLM agent harnesses, run against ARC-AGI-3. Change one piece of the scaffold at a time and measure whether the same model solves more.

GameNight

gamenight.yaqzan.dev ↗

Tier list of my board game shelf. Adding a game kicks off an agent pipeline that researches and rates it, and the whole UI renders from one validated JSON file.

All projects →

Skills

Languages
Ruby · Python · JavaScript · TypeScript · SQL · Bash / PowerShell
Front end
React · Vue.js · Tailwind CSS · HTML5 · CSS
Back end
Ruby on Rails · Django · Flask · REST · GraphQL · Sidekiq · Elasticsearch
Cloud & infra
AWS (EC2, RDS, S3, ACM, IAM, CloudWatch) · GCP · Azure · Terraform · Pulumi · Docker · Nginx · GitLab CI
Data
PostgreSQL · MySQL · SQLite · Elasticsearch
Testing
RSpec · PyTest · Jest · Capybara · Playwright · Selenium
Practices
SOC 2 · Auth & SSO (OAuth, Okta/SCIM) · Agentic AI development (Claude Code) · Mentorship

Education

Western University

BSc, Specialization in Computer Science

2016