Selected moments with collaborators, colleagues, and teams across research and industry.
A small studio for serious AI systems
TensorHarmony is a Canadian applied-AI studio focused on applied research, intelligent products, and production systems. We build our own products, partner with founders, and work with clinical-AI teams on the hard parts: computer vision, infrastructure, validation, and deployable software.
How we work
The studio operates on a small set of working principles. They show up in every engagement and every product we ship.
Build to ship
A model in a notebook is not a product. Every engagement is shaped to end in something a real user, a real team, or a real reviewer can actually pick up and use.
Research-grade engineering
Reproducibility, evaluation discipline, and honest documentation are not optional extras. They are how serious AI products are built, even when the domain is not regulated.
Full-stack ownership
Models, APIs, mobile, cloud, payments, infrastructure, and the boring bits in between. We are comfortable owning the whole vertical when that is what the work needs.
Honest about trade-offs
We tell clients what we believe, including when an idea is wrong, when an approach is not ready, or when someone else is the better fit.
Long-term thinking
We prefer relationships that compound: products we operate for years, partners we ship multiple things with, and codebases we would be proud to return to in three years.
Built for collaboration
TensorHarmony is intentionally small and designed to grow carefully through co-founder relationships, joint products, senior collaborators, and focused partnerships on the right problems.
The team
A small, principal-led studio with a senior bench
Founder and principal engineer at TensorHarmony, currently serving as Principal Machine Learning Engineer for GuideAI Health's vascular-AI platform and first FDA submission. Previously spent nearly seven years at IBM Watson Health, mostly on the Liver Advisor program, contributing to three production CT/MRI releases and co-invented US patents on lesion detection. Founder of FaceHarmony, a shipped consumer generative-AI product for iOS and Android. Comfortable across the stack from CUDA and ML pipelines to cloud, mobile, and production operations.
Assistant Professor of Computer Science at Nipissing University with expertise in computer graphics, visual computing, image stylization, and vision-based processing. Former Graphics Engineer at Huawei Technologies' HMI Lab and Ph.D. graduate from Carleton University. Collaborates with TensorHarmony on selected graphics, visualization, and generative-imaging initiatives.
Proudly Canadian
Based in Richmond Hill, Ontario, Canada
Open to collaboration
Building together
TensorHarmony is actively interested in a few kinds of conversations beyond standard paid engagements:
Technical co-founder conversations
For founders with a real wedge in a market we understand — applied AI tooling, computer vision, vertical SaaS, medical imaging, or production AI infrastructure — we are open to co-founder-shaped conversations on the right problem.
SaaS and product partnerships
Joint product builds where TensorHarmony brings applied AI, full-stack engineering, and production discipline while a partner brings domain knowledge, distribution, capital, or customer access. Equity-bearing structures are possible for the right fit.
Senior collaborators
The studio occasionally works with senior researchers, engineers, designers, and domain specialists on specific projects. If your work overlaps and you want to ship something real together, get in touch.
Let’s Build Something Deployable
For founders, research teams, and product builders looking for applied AI engineering, technical partnership, or production-focused collaboration.
Founder, partnership, and technical collaboration conversations welcome