About Momentaj
Twenty years of systems that aren't allowed to fail, now pointed at AI.
Who we are
Momentaj is a Toronto AI engineering firm, founded in 2013 and led by its founder, Amir Pournasserian. We design and build AI agents, MCP servers, retrieval systems and AI-native products, and the software, cloud and integration work those systems depend on. We work remotely with clients across Canada, the United States, the Middle East and beyond, and in person across the Greater Toronto Area.
The company's history is in systems with no tolerance for failure: a real-time tracking system for Canada's Department of National Defence, a national online travel agency's booking platform, a lending ledger that holds other people's money, and IoT platforms ingesting a hundred million readings a week. Since 2024 that same discipline has gone into AI, and that is where most of our work is today.
How we work
Small and senior. A core team of up to ten people, extended per engagement with specialists we have worked with before. The founder is hands-on on every project: architecture, code review, and usually code.
Thirteen years, one thread
Take on the system nobody can afford to get wrong, and build it so it doesn't. The technology underneath has changed more than once; the job hasn't.
-
2013
Momentaj founded in Toronto
Software development and consulting, with a first decade in platforms, payments and IoT.
-
2017
uBeac and the research programme
Founded uBeac, a multi-tenant IoT platform that grew to 3,000+ users, and began indoor-positioning research with George Brown College.
-
2019
Federal R&D contract
Won a competitive Innovative Solutions Canada contract from the Department of National Defence and delivered Momentrack with a team of nine, on schedule.
-
2019 to 2022
Trip Support
Fractional CTO and chief architect for a Canadian online travel agency: one booking platform, fifteen integrations, a team grown from one person to fourteen.
-
2022 to 2023
Private Lending Platform
Architected and built a lending platform that funds instalment plans with investor capital, now in production.
-
2023
FluentCMS and YeSvelte
Released two open-source projects, since grown to 780+ GitHub stars between them.
-
2024 to 2025
AI platforms
DevGuardian AI (multi-agent code review), Valco AI (real-estate intelligence) and MayAI (a consumer agent) designed and built.
-
2026
MCP and AI tooling research
Published comparisons of MCP servers for websites, MCP testing tools and the rules each AI host applies, used in our own client work.
Founder
Amir Pournasserian
Founder & CEO
Education
- BSc, Electrical & Electronics Engineering, Shahid Beheshti University, 1999
- Master's, IT Project Management, Industrial Management Institute, 2005
- Data Science Specialization, Johns Hopkins University, 2014 to 2017
Amir started as an electrical engineer, but the systems that pulled him in were always digital. In 2002 he founded Tajan, a banking software company he grew from two hired developers to fifty people over fifteen years as CEO. Its products, centralised identity and single sign-on, transaction reconciliation across more than fifty financial formats, and nationwide ATM monitoring, served twelve major banks under ISO/IEC 27001 certification. Tajan was acquired in 2021.
He founded Momentaj in 2013. Through it he won a competitive Innovative Solutions Canada contract from the Department of National Defence and led the nine-person team that delivered Momentrack; served as fractional CTO and chief architect for Trip Support, where he designed its booking platform and grew its team from one person to fourteen; and architected a private lending platform that is in production today. Alongside client work he founded uBeac, an IoT platform acquired in 2020, co-founded the hardware startup Tingslab, and ran a research programme with George Brown College, Ryerson University (now Toronto Metropolitan University) and the University of Waterloo.
The last few years have been about pointing that same discipline at AI: DevGuardian AI, a multi-agent code review platform; Valco AI, a real-estate intelligence platform; and MayAI, a consumer agent. Today he is Principal AI Engineer/Architect at Four Seasons Hotels and Resorts, where he leads AI development for its AI-powered hotel concierge, discovery and booking platform, a multi-agent system on .NET, Microsoft Agent Framework and the Model Context Protocol that reaches guests through ChatGPT and Claude. He created FluentCMS and YeSvelte, and publishes research on MCP tooling and AI host requirements at pournasserian.com.
Research and collaborations
Alongside client work, Momentaj has run a research programme since 2017 on the hard problems underneath the products we build: radios and power budgets, data pipelines at scale, indoor location, and how far a web framework can be pushed.
- George Brown College. An intellectual-property MOU (October 2017) with the Office of Research & Innovation for "Evaluation of Beacons and Gateways for Indoor Positioning Systems", under the college's NSERC Industrial Research Chair in Smart Connected Buildings, with two student researchers.
- Ryerson University (now Toronto Metropolitan University). A joint NSERC Engage proposal (January 2018) with the Mechanical & Industrial Engineering department to improve beacon location accuracy from about 6 m to 2 to 3 m.
- University of Waterloo. A co-op student joined the Momentrack hardware and firmware team from January to April 2020.
- Indoor positioning. Four rounds of experiments across five gateway brands and nine beacon types; BLE fingerprinting with a random forest reached 85% zone accuracy.
- Blazor rendering. Runtime compilation and renderer-level caching in FluentCMS cut a page render from 45 ms to 3.8 ms and lifted an 8-core server from about 250 to about 1,120 concurrent users.
What we hold ourselves to
Production first
A demo is not a deliverable. We design for the day the system is running without us watching it.
Hands-on, founder-led
The person who scopes your project is the person who architects it, reviews the code and answers for it.
Honest about AI
We will tell you when a model is the wrong tool, when your data isn't ready, and what a wrong answer will cost.
Open by default
We publish our research and our open-source projects, so you can check how we think before you hire us.
Let's talk about your AI project
Tell us what you are trying to do. We will come back with an honest take on what it would take.
Book an AI consultation