AI engineering · Toronto
AI systems that hold up in production
Momentaj designs and builds AI agents, MCP servers and retrieval systems for companies that can't afford to get them wrong. Founder-led, hands-on, and backed by twenty years of banking, travel and defence systems.
Clients, partners and research collaborators
What we build
Six ways we put AI to work, each backed by something we have shipped.
AI Agents & Multi-Agent Systems
Role-based agents that act on real systems: tool calling, memory, structured replies and human approval where it matters. Built on Microsoft Agent Framework, LangChain and OpenAI-compatible models.
Learn moreMCP Servers & AI Integrations
Put your product or data in front of ChatGPT, Claude and other AI hosts through the Model Context Protocol. Secured with OAuth 2.1 and tested against each host's rules.
Learn moreRAG & Knowledge Systems
Retrieval built for your material, not generic chunking: hybrid search, GraphRAG, code-aware indexing and incremental refresh on pgvector, Azure AI Search or Neo4j.
Learn moreAI-Native Product Development
A product designed around the model from day one, with the platform engineering to run it: multi-tenancy, queues, observability and cost control.
Learn moreAI Strategy & Readiness
A short, hands-on assessment of where AI pays off in your operation, what data and guardrails you need, and a build plan with honest estimates.
Learn moreApplied Machine Learning & Forecasting
Forecasting, scoring, clustering and anomaly detection on your own data, in Python on PyTorch and TensorFlow, with the data pipelines that make the models trustworthy.
Learn moreRecent AI work
Agents, retrieval and machine learning we have designed and built, with the numbers behind each one.
DevGuardian AI - Multi-Agent Code Review Platform
Valco AI - Real Estate Intelligence Platform for Dubai
MayAI - A Consumer AI Agent for Household Tasks
Production-grade by design
A demo that impresses and a system that runs are different products. Every AI system we ship has these four things built in.
Grounded
The model returns IDs; the server builds what the user sees. A product, price or action the model invented never reaches the screen.
Secured
OAuth 2.1 and Entra for MCP servers, role-based access, PIPEDA-aware data handling, and our founder's fifteen years of banking software, at a company certified to ISO/IEC 27001 from 2016.
Observable
OpenTelemetry and Application Insights from day one. We measure every hop, so when something degrades we know which one.
Evaluated
Evals and MCP conformance checks before launch, with cost and latency budgets per tool call, so the prototype that impressed still works at month six.
How we work
Short steps with something real at the end of each one, so you can stop or change direction at any point.
Discover
Interviews with the people doing the work, a look at your data and systems, and a written plan: what to build first, what it needs, what it would cost.
Prototype
A working prototype (a proof of concept) on your real data: an agent, MCP server or retrieval system, built to find out what breaks before it matters.
Production
Security, evaluations, observability and rollout. The prototype becomes a system your team can run and your auditors can read.
Operate
Monitoring, model and cost updates, and the next capability, with the same engineers who built it.
Insights
Research notes and working patterns from building agents, MCP servers and retrieval systems.
October 8, 2026 · 4 min read
Your website has a new visitor: AI agents
An MCP server for a website lets AI agents search the site, ask it questions and act on it for a person, and who runs that server decides what agents see. A summary of founder Amir Pournasserian's October 2026 survey of 44 tools, and what it means for a business.
October 8, 2026 · 4 min read
Should your product have an MCP server? A decision guide
An MCP server lets AI agents in ChatGPT, Claude and other hosts use your product through a small set of tools. This decision guide for SaaS and platform teams gives five signals to build one now, three to wait, and what a good first server looks like.
October 8, 2026 · 4 min read
Keeping AI agents honest: the server-side grounding pattern
Server-side grounding is a pattern in which an AI agent returns only identifiers and the server builds everything the user sees from real tool results, so the agent cannot show an invented item. How we built it for MayAI, the six rules around it, and where it applies.
Built in the open
Three projects we created. Read the code before you hire us.
FluentCMS
★ 560+An open-source CMS for .NET, built on ASP.NET Core and Blazor, that works as both a page-building CMS and a headless CMS on one API, with an AI writing assistant built in.
YeSvelte
★ 222Svelte components whose look is a swappable theme: Tabler (Bootstrap) or daisyUI (Tailwind). Switch themes by swapping one stylesheet.
uBeac
The backend of an IoT platform that turned off-the-shelf Bluetooth gateways into live building dashboards: ingest, queue, decode, store, stream. Acquired in 2020 and published as a reference architecture.
Founder-led
Amir Pournasserian
Founder & CEO
Amir has spent more than twenty years on systems that aren't allowed to fail: banking platforms for twelve banks, a real-time tracking system for Canada's Department of National Defence, and a national travel booking platform. Since 2024 that discipline has gone into AI: multi-agent code review, real-estate intelligence, a consumer agent, and today the guest-facing AI platform of Four Seasons Hotels and Resorts, where he is Principal AI Engineer/Architect.
The engineering underneath
Most of what an AI system does is call ordinary software. We build that too, and we build it to last.
Custom Software Development
Platforms, back offices and integrations built to run for years: .NET and Python services, queues, payments, real-time delivery, and the APIs an AI system will need.
Cloud & DevOps
Azure and AWS infrastructure for AI and platform workloads: identity, API management, observability, CI/CD and cost control, as code.
IoT & Connected Systems
Devices, radios and the cloud that receives their data: BLE, LoRa and LTE-M hardware, MQTT ingestion, real-time dashboards and the analytics on top.
Web & Mobile Development
Fast, accessible web apps and mobile-first PWAs in Blazor, SvelteKit and React, designed with the people who will use them.
Have an AI project,
or not sure yet?
Tell us what you are trying to do. We will come back with an honest take on what AI can and can't do for it, and what it would take to build.