I'm Darian Lagman, a full-stack engineer based in Winnipeg, Canada, finishing an Honours Computer Science degree at the University of Manitoba with minors in Statistics and Physics & Astronomy.

I work at the seam between AI-assisted product development and the systems those products run on — which means I ship fast with Claude Code and Codex, and I can also defend the math, the database choices, and the auth model when someone asks. Right now most of my time goes into Stockman, a SaaS inventory and decision-support platform with an LLM/MCP assistant layer, and a from-scratch Holt-Winters forecasting engine that powers it.

I'm looking for a full-time engineering role where the team takes both speed and depth seriously.

The longer version

I started programming because I wanted to simulate things — orbits, populations, anything where the rules were simple but the behavior wasn't. That's still the through-line. The N-body simulator I built for my Physics & Astronomy minor and the forecasting engine I built for Stockman are the same instinct: take a system that looks chaotic, decompose it into level / trend / seasonality (or position / velocity / acceleration), and find out what's predictable and what isn't.

The full-stack work is the same instinct applied sideways. Stockman started because a small distributor was running their entire business out of CSV files, and the gap between "spreadsheet that works" and "system that scales" is mostly engineering, not magic. React on the front end, Spring Boot and MySQL on the back, session-based auth with role-based access because the product is a back-office tool and stateless tokens were solving a problem we didn't have, Docker Compose for dev parity, and an MCP layer so the assistant can call the same APIs the UI does, with the same permissions.

I use AI tooling — Claude Code, Codex, Cursor, and a stack of local models running on a homelab box — the way a senior engineer uses any other force multiplier: aggressively, with skepticism, and with a hand-tuned AGENTS.md so the model has the same constraints I'd give a junior. The point isn't speed for its own sake. The point is that I can ship a feature today and six months from now still know exactly why every line is the way it is.

Outside of work I fish in Manitoba lakes, garden, and read more astronomy than is strictly justified. I'm currently looking for a full-time role; my email's at the bottom of every page.

What I'm optimizing for

• Technical depth — math, systems, and code that holds up under review.
• Shipped impact — products real users touch, not slide-deck demos.
• The right team — small, sharp, willing to read papers and write tests.

Where I am

Winnipeg, Canada. Open to remote, hybrid, or relocation. See /now for what I'm currently working on, or /resume for the formal version.