Elias TanbeARTIFICIAL INTELLIGENCE ARCHITECT.

AI ARCHITECT · MONTRÉAL, QUÉBEC

Elias Tanbe.

Artificial Intelligence
Architect.

I help companies in Québec and across Canada explore AI in three stages, with people kept in control.

Take the 1-minute AI check
Visit the colony ↘Book a call ↗The three stages ↓
WHY ANTS?Step into the colony No single ant builds a colony. Come see how I build AI.

FREE · 1 MINUTE · 5 QUESTIONS

Where should your company
start with AI?

Answer five questions and see which of the three stages fits you today: explore, pilot or deploy. Your answers stay in your browser.

  1. 01 · TodayWhere is your company with AI today?
  2. 02 · The workIs there a task that takes your team hours every week?
  3. 03 · The dataAre the documents or data behind that work easy to reach?
  4. 04 · PrivacyDo you know which personal information that work involves, and who is responsible for it?
  5. 05 · The peopleIs someone ready to own a trial?

SERVICES / THREE STAGES

Explore AI
in three stages.

Most companies don’t need another AI strategy deck. They need to know where AI helps their work, prove it on one real case, then make it part of daily work. Each stage is quoted on its own, quickly and in writing, and ends with a decision that stays yours: continue, adjust or stop.

  1. 01Explore

    Where can AI help, and where shouldn’t it?

    We map your processes, data and constraints with the people who do the work. I check your privacy obligations under Québec’s Law 25 before any data moves.

    You leave withUse cases ranked by value, risk and effort, and a clear recommendation to go ahead or not.

    In practiceOYEE helps people understand AI through questions and examples, in French and English, with in-person workshops in Québec.

  2. 02Pilot

    Prove one use case on your real data.

    We agree on success criteria before anything is built. I build a working pilot where people review and approve what the AI produces, then measure it against how you work today.

    You leave withA working pilot, measured results and an honest recommendation.

    In practiceHaulView · Iris prepares each offer and the dispatcher approves it. In development; client details stay confidential.
    Adra a working prototype built solo in five hours, top 4 of about 30 teams at Mila Ventures’ Build OS26.

  3. 03Deploy

    Put it into daily work, safely.

    I take the pilot to production with security, monitoring, cost control and the human checkpoints that remain. Your team is trained to run it.

    You leave withA system in production, its documentation and a team that can run it.

    In practiceMnemosyne a bilingual learning product in production, used by real learners.

French and English
Workshops, documents and the AI itself in French, English or both.
Privacy by design
Law 25 and Canadian privacy requirements shape the design from the first stage.
People stay in control
AI prepares the work. A person approves wherever a decision matters.
Built, not only advised
I design, build and run AI systems in production, in education and in operations.
Read the guide: starting with AI in a Québec company ↗See how I build ↓
Elias Tanbe

ABOUT / WHO YOU WORK WITH

I build AI systems
where people stay in control.

I’m an AI architect in Montréal, and I write the code myself. I design, build and run AI systems: learning products used by real learners, operations tools where a person approves what the AI prepares, and multi-agent reviews. I founded Inquisitive Flow Learning, and my experience as an educator shapes how I bring teams along.

  • Founder, Inquisitive Flow Learning
  • Top 4 of about 30 teams, Build OS26, Mila Ventures
  • French · English · Arabic

What I bring

  • Agentic and multi-agent systemsClear roles, handoffs and checks between agents, with a person approving what matters.
  • Generative AI productsFrom prototype to production: memory, verification and an interface people actually use.
  • Evaluation and verificationModels tested on your own examples, and what the AI says checked before anyone relies on it.
  • Privacy and governancePrivacy by design and Québec’s Law 25, from the first stage.
  • Teaching and adoptionAn educator’s way of bringing teams along, and AI literacy for people new to it.

Tools and infrastructure

Infrastructure
AWS · DigitalOcean · Cloudflare · Google Cloud
Pre-AI markup and coding knowledge
HTML · CSS · Bootstrap · Java · Python · PHP, all coded line by line
Expanded with AI
Swift · JavaScript · TypeScript · React · FastAPI · Node.js · Three.js · SQL
Git and GitHub
Branching · Pull requests · Code review · Merge and rebase · CI/CD with GitHub Actions · Dependabot

Across my products: PostgreSQL · Redis

How I work ↗LinkedIn ↗

WHAT I’VE BUILT

Products, live and in progress

Also built

See all my work in the colony ↘

RESEARCH / ON MY OWN TIME

Two experiments,
in progress.

Beside my client work, I run two research experiments on my own time. Both follow an academic approach: a clear question, a hypothesis, a test, and results I can defend. They are in progress, so no results are published yet.

  1. 01Memory

    How do people remember?

    I study how people remember what they learn: what helps knowledge last, what lets it fade, and what that means for AI built to support learning.

    StatusIn progress. No results published yet.

  2. 02Hypotheses

    Can AI form a hypothesis, then test it?

    I study whether an AI system can work the way a researcher does: form a hypothesis, decide how to evaluate it, then test it in code to confirm or reject it. The aim is an academic method, not a confident guess.

    StatusIn progress. No results published yet.

Also exploring

Working on memory, hypothesis testing or coding agents? I like comparing notes with other builders and researchers.

Connect with me on LinkedIn ↗or write to me: elias.t@inquisitiveflow.ca

QUESTIONS

Questions companies
ask first.

What does an AI architect do?

An AI architect decides how AI fits into a company’s work: which tasks it takes on, which data it may use, how its output is checked and where a person decides. I design that system, build it and make sure it runs safely at a known cost.

What is the difference between an AI architect and an AI engineer?

The architect decides how the whole system fits the work: what AI takes on, which data it may use, how its output is checked, what it costs and where a person decides. AI engineering is building it: the code, the integrations, the tests and the production setup. I design the system and write the code myself, so the design is proven in working software rather than left in a slide deck.

What AI projects have you built?

Mnemosyne, a bilingual learning product in production with real learners, with memory and a layer that verifies what the tutor says. HaulView, an operations product in development where its AI layer, Iris, prepares each offer and a dispatcher approves it. Adra, a multi-agent review prototype built solo in five hours, top 4 of about 30 teams at Mila Ventures’ Build OS26. And OYEE, a bilingual way to understand AI, with workshops in Québec. I also built learning products and free tools for educators at Inquisitive Flow Learning.

Where should a company start with AI?

With its own work, not with a tool. In the Explore stage we list the tasks where AI could help, rank them by value, risk and effort, and choose one to prove in a pilot.

How much does it cost?

It depends on the stage and the size of the problem, so I don’t publish fixed prices. After a 30-minute call you get a quick written quote for the stage you need, before any commitment.

How do you handle privacy and Québec’s Law 25?

From the first stage. We identify the personal information involved, keep only what is needed and complete a privacy impact assessment when the law requires one. When data must stay in Canada, the architecture and the AI providers are chosen accordingly. For legal opinions, I work with your counsel.

Do you work in French?

Yes. I work in French and English, and I also speak Arabic. Workshops, documentation and the AI systems themselves can be in French, English or both.

Do you only work in Montréal?

I’m based in Montréal and work with companies across Québec and the rest of Canada, on site or remotely.

What do we need before we start?

A problem worth solving, someone who knows the work well, and examples of the documents or data involved. You don’t need a data science team to begin.

What happens after the pilot?

You decide. If the pilot meets the criteria we set together, we move to deployment. If it doesn’t, you keep the findings and we adjust or stop. A pilot that says no is also a useful result.

Which AI models do you use?

The ones that fit your constraints: quality in your language, cost, privacy and where data may be processed. I compare options on your own examples instead of choosing a vendor in advance.

Which technologies do you build with?

Python, TypeScript, JavaScript, Swift and SQL, with FastAPI, React, Node.js and Three.js, PostgreSQL and Redis, on AWS, DigitalOcean, Cloudflare and Google Cloud. The work goes through Git and GitHub, with code review and CI/CD.

AI architect or AI engineer? Read the full comparison ↗

HOW I BUILD / WHY A COLONY?

No single ant
builds a colony.

The larger structure emerges from small actions, shared signals and different roles. That is how I build agentic systems: one agent gathers context, another challenges it, another brings the work together. I design those relationships, the checks between them and the moments where a person must decide.

Ants Agents
Each one has a clear role.
Tunnels Handoffs
Context travels from one agent to the next.
Chambers Decisions
Where the work is shaped, checked or approved.
The colony The system
Larger and more reliable than any single agent.

THE COLONY / SIDE VIEW

Four routes.
Twenty-four chambers.

Each route is one part of how I build. Each chamber is one decision inside it. Choose a chamber to step inside.

  1. 01Frame the problem4 chambers
  2. 02Orchestrate the work7 chambers
  3. 03Verify and authorize5 chambers
  4. 04Ship and observe8 chambers
A LIVING ARCHITECTURE / SIDE ELEVATION 4 chambers 7 chambers 5 chambers 8 chambers 01 Root amphitheatre 02 Exposed roots 03 The oculus 04 First crossing 01 Handoff docks 02 Woven gallery 03 The confluence 04 Memory archive 05 Dispatch galleries 06 A second passage 07 The joining loom 01 Hall of evidence 02 The review chamber 03 Human threshold 04 The open crescent 05 The boundary 01 The load-bearing hall 02 Water reservoir 03 Finishing workshop 04 The growing garden 05 Evidence archive 06 Release terraces 07 The observatory 08 Living root sanctuary SURFACE DEEPER ↓
Enter the colonyRead it as text ↗

START WITH A CONVERSATION

What would you like
AI to help with?

A 30-minute call is enough to see whether AI fits and which stage to start with. You then get a quick written quote, before any commitment. In French or English.

Book a 30-minute callConnect on LinkedIn ↗Not sure yet? Take the 1-minute AI check ↑

Or write to me directly: elias.t@inquisitiveflow.ca

Building with AI too?

I like comparing notes with other builders and researchers, on my two experiments in progress or on what you are building.

See my research ↑Connect with me on LinkedIn ↗

Explore Elias Tanbe’s colony

Finding a way beneath the surface…

WHY A COLONY?

The way I work.
Beneath the surface.

The colony makes coordination visible. In my agentic systems, the connections are explicit: typed handoffs, shared state, parallel work, verification and human approval. Explore the engineering decisions behind the agents.

A LIVING ARCHITECTURE0 OF 24 CHAMBERS EXPLORED

THE COLONY / ALL 24 CHAMBERS

A world beneath the surface.

A side view of the whole colony. Choose any chamber to step inside.