Vozwin AI is the AI consulting division of Vozwin Inc., a multi-divisional engineering firm founded in Montreal in 2015. We are a bilingual team based in Pointe-Claire, Quebec, working with companies across Canada to assess where AI fits, build the systems that make it work, and train the people who will run them. The team has more than 30 years of experience bringing new technology to market, and our advice is backed by Vozwin's own peer-reviewed machine learning research with McGill University and Université de Sherbrooke.
This page covers who we are, how Vozwin AI fits within the Vozwin group, the research behind our work, who we work with, how an engagement runs, and the principles we hold ourselves to.
Who we are
Vozwin AI is run by the Vozwin Team, the same group behind Vozwin's engineering, aerospace and research work. Collectively, we have spent more than 30 years launching products, building companies and leading the technical teams that ship them. That background shapes how we consult: we judge AI by what it changes in your operation, not by how impressive it looks in a demo.
We work in English and French from our office in Pointe-Claire, on Montreal's West Island. The Vozwin Team signs our guides and stands behind every recommendation we make, and when a guide cites a figure, it links to the source.
Part of the Vozwin group
Vozwin Inc. has three divisions, AI, Aerospace and Labs, plus a venture studio that takes new technology from idea to company. Each does different work, and having them under one roof is what keeps our AI advice practical.
- Vozwin AI: AI consulting, from readiness assessments and strategy to implementation, training and change management.
- Vozwin Aerospace: aircraft design, analysis and integration, including UAV platforms that log their own operating data from the first flight.
- Vozwin Labs: applied research and development and design engineering, including the machine learning research described below.
- Venture Studio: takes new technology from idea to launch. Vermilion, a physics-informed reasoning model built on Labs research, was incubated there and now operates as its own company.
For you, this means the people advising you on AI work alongside engineers who design hardware and researchers who publish. When a project touches sensors, embedded systems or physical equipment, we understand it from the inside rather than bringing in another firm to translate.
The research behind our advice
Vozwin funds and owns applied machine learning research through its Labs division. Under the PHUMS project, Vozwin worked with McGill University and Université de Sherbrooke, supported by the Mitacs Accelerate program, to predict drone battery state of health from flight logs. The peer-reviewed result predicted battery health within 2.26 percent on a battery the model had never seen, using 631 flights and only the voltage, current and throttle a stock flight controller records. Continued training on more data has since brought the error down to 1.57 percent.
We mention it because it changes the advice we give. When we tell you whether your data can support a model, how many examples you need, or whether a vendor's accuracy claim holds up, we are drawing on a project we took from data collection to published result with a small dataset and real constraints. Our guide on drone battery state of health walks through the full method.
Who we work with
We work with growing businesses that need to move fast and large organizations working through complex change, in sectors including aerospace, energy, oil and gas, manufacturing, healthcare, finance, retail, hospitality and technology. The industry changes, but the questions stay the same: what decision should change, what data supports it, and who will act on it.
We serve companies across Canada, and we scope projects with Canadian funding in mind. SR&ED tax credits, NRC IRAP and Mitacs can offset a meaningful share of eligible AI work when the project is designed around them from the start.
How we work
Every engagement follows the same four steps, sized to the problem in front of us:
- Discover. An honest conversation about your goals, your constraints and what is realistic. No sales deck.
- Assess. Where AI fits your business, what your data and systems can support, and what your team is ready for.
- Build. Strategy, vendor selection, architecture and shipped implementations, with pilots measured against a baseline before any big bet.
- Embed. Training, governance and the operating model that keep the capability inside your company after we leave.
We are vendor-neutral. For each use case we review the market, test shortlisted products on your data, and recommend a custom build only when no product fits or the data advantage is one you should own. That includes Vermilion: when it is the right platform we say so, and when something else fits better we say that instead.
What we hold ourselves to
- Outcomes over hype. We are not here to sell AI for its own sake. If a simpler tool or a process change solves the problem, we will tell you.
- Honest guidance. We tell you what you need to hear, including when a project is not ready or not worth funding yet.
- Strategy before tactics. AI has to fit your wider business plan, so we start there rather than with the latest model.
- Capability stays with you. We train your people and document what we build so you are not dependent on us after the engagement.