About
Amin Mirlohi
Amin Mirlohi is an AI engineer and independent builder with a PhD in Computer Science, working on retrieval systems (RAG), multi-agent orchestration, and the evaluation of AI systems, and writing publicly about the engineering behind them.
I work on the systems that make AI useful in the real world: retrieval pipelines that ground answers in real sources, multi-agent systems that plan and act without falling over, and the evaluation harnesses that tell you whether any of it actually works. I build products and think like a founder, and I write about the craft as I go.
- Role
- AI Engineer & Independent Builder
- Based in
- Toronto, Canada
- Focus
- AI for learning, Retrieval systems, Evaluation
What I focus on
My work centers on three things: retrieval-augmented generation, multi-agent systems, and machine-learning evaluation, increasingly applied to AI for learning. The common thread is grounding and measurement. It is easy to make a demo that looks intelligent; the hard, interesting part is building systems that stay correct, cite their sources, and degrade gracefully when they hit the edges of what they know.
Background & research
I hold a PhD in Computer Science from the University of New Brunswick. My research spans information retrieval, network analysis, adversarial & applied machine learning. That work shapes how I think about modern retrieval and AI systems today. Over the past decade I have built AI products and led technical teams across both research and industry, taking ideas from prototype to production.
- PhD, Computer ScienceUniversity of New Brunswick · 2016–2020
- MSc, Computer Engineering (Artificial Intelligence)Sharif University of Technology · 2012–2015
- BSc, Computer Engineering (Software Engineering)University of Isfahan · 2007–2012
Currently at Ametros (education technology).
How I think about building
Evaluation comes first. Before I trust a system, I want a dataset, a metric, and a way to catch regressions; otherwise “it works” is just a feeling. I favor deterministic guards over hopeful prompting, retrieval over recall-from-weights when facts matter, and small, legible components over monoliths that are impressive until they break. Good engineering judgment, not model hype, is what makes AI dependable.
Writing
The clearest picture of how I think is in the writing. Articles work through the engineering ideas in depth; the journal is a running log of what I am building and learning.
Frequently asked
- Who is Amin Mirlohi?
- Amin Mirlohi is an AI engineer and independent builder with a PhD in Computer Science. He works on retrieval systems (RAG), multi-agent orchestration, and the evaluation of AI systems, and writes publicly about the engineering behind them.
- What does Amin Mirlohi work on?
- His work centers on retrieval-augmented generation, multi-agent systems, and machine-learning evaluation, with an applied focus on AI for learning. He builds production AI systems and writes long-form articles and a building-in-public journal about the craft.
- What is Amin Mirlohi's background?
- He holds a PhD in Computer Science from the University of New Brunswick, an MSc in Artificial Intelligence from Sharif University of Technology, and a BSc in Software Engineering from the University of Isfahan. His research spans information retrieval, network analysis, and adversarial/applied machine learning, and he has years of experience building AI products and leading technical teams.
- Is Amin Mirlohi available for consulting?
- This site is Amin's personal home as an independent builder, not a consulting service. The best way to follow his work is through the Articles and Journal here; for direct correspondence, use the contact page.
- Where can I read Amin Mirlohi's writing?
- His long-form Articles and short building-in-public Journal entries are published on mirlohi.com, covering retrieval systems, multi-agent design, evaluation, and applied machine learning.
- How can I contact Amin Mirlohi?
- By email at amin@mirlohi.com, or via LinkedIn and GitHub. Links are on the contact page.