Reel 01 · Waterloo · 2026

MahyarTajeri

I build fast software. Sometimes it controls knives.

Satellite-network simulators in Rust. Vision models that hold up outside the lab. The brains of a machine that slices bread to the millimeter. I go looking for problems that bite back.

Rust · Machine Learning · Things That Run Fast

See the Work Let's Build

Scroll
Scene I — The Operator

About

Speed and precision. That's pretty much the whole pitch — I like the problems where the math is actually hard and "close enough" doesn't get you paid.

I started out writing weather code for the federal government. These days I'm at Waterloo working out how to route data between satellites moving at 7 km/s. Somewhere in the middle I wrote neural nets by hand, trained vision models, and programmed a machine that cuts bread better than I can.

Software, math, hardware — I don't really have a favorite. I have one question: can I make it work, and can I make it fast. So far the answer keeps being yes.

Now: Backend @ Abnormal AI
Researching: LEO satellite networks @ Waterloo (MMath)
Status: Building. Always.
Scene II — The Evidence

Selected Work

One iteration closer to freedom — House Rule
Scene III — The Track Record

Experience

May 2026 — Present

Backend Developer · Abnormal AI

Keeping flagship services fast when the traffic gets ugly.

  • Built a predictive auto-scaler on Fourier-series regression — cut prediction error to a third of the old heuristic.
  • Reviews, fixes, and docs, all in a fully AI-native workflow.

Jan 2026 — Present

Research Assistant · University of Manitoba

The satellite-network simulator up above started here.

  • Built in Rust for raw speed and bit-for-bit reproducibility, with interactive 3-D visualization.
  • Benchmarking latency, throughput, and resiliency across classical and ML-based routing.

May 2023 — Aug 2024

Software Developer · Environment & Climate Change Canada

Code that millions of Canadians actually open.

  • Shipped features for weather.gc.ca — new endpoints and data visualizations in Vue.
  • Wrangled big data into something usable; ran remote servers and MySQL/Postgres databases.
  • Built internal Python tooling and reviewed PRs with senior engineers.

May 2022 — Sep 2022

Data Analyst · Ministry of Public & Business Service Delivery

Turning raw infrastructure data into something people could act on.

  • Built Power BI reporting and piped data through Azure Data Factory into SQL servers.
  • Reviewed PowerShell tooling and tracked fleet health from the terminal.
Scene IV — The Arsenal

Skills

Languages

  • Rust
  • Python
  • C++
  • Java
  • JavaScript
  • SQL
  • HTML / CSS

Machine Learning

  • PyTorch
  • TensorFlow
  • Nets from scratch
  • Computer vision
  • Deep RL
  • Numerical methods

Systems & Cloud

  • REST APIs
  • Vue & React
  • Flutter
  • AWS & Firebase
  • MySQL / Postgres
  • MongoDB
Scene V — The Invitation

Let's Build Something

Got something ambitious, ugly, or a little bit unreasonable? That's usually the good stuff. Drop me a line.