Lemay.ai is a Canadian AI engineering firm registered to contract in the United States. We design, build, deploy, and govern AI systems for defense programs, secure enterprise environments, and federal and state agencies, as a prime where scope allows, and as a subcontractor to primes who need AI capability they do not want to build in-house.
Because the constraint is rarely geography: it is finding a supplier who has actually put AI into production in a regulated or classified environment, and who will still be reachable in year three. We have delivered 250+ projects since incorporating in 2016, across 21+ countries, in aerospace and defense, government, energy, financial services, and healthcare.
Practically: our U.S. registration is current (Unique Entity ID JJADR6LWJ743), we hold NATO codification (NCAGE L0QB4), we are a member of the SOSSEC consortium for U.S. government teaming, we are registered under Canada's Controlled Goods Program, and members of our team hold Top Secret clearance. We won the NATO Defence Innovation Challenge in 2020 with Reticle.
Commercially: fixed-fee work plans, full client ownership of project-specific IP, founder-led delivery, and knowledge transfer at handover so your team owns what we build. No hourly creep, and no dependency on us after the engagement closes.
Six capability areas, each with a page that goes into detail.
Anti-drone and counter-UAS classification, sensor fusion across radar and EO/IR, ISR and OSINT workflows, predictive maintenance for fleet readiness. Defense AI engineering →
Generative AI and retrieval systems deployed inside your own network, where the document set and the query content never leave. Secure GenAI deployment →
Self-hosted inference, offline installers, and edge deployment for environments with no outbound internet access. On-prem AI consulting →
The unglamorous part: runtime selection, MLOps, monitoring, drift detection, retraining, and the cost and latency engineering that decides whether a system is affordable to run. Deployment & MLOps →
Contract research where the method does not exist yet or the physical experiment is too expensive to run blind. 4 PhDs, 90+ peer-reviewed publications. AI R&D partner →
ISO/IEC 42001:2023 implementation and certification readiness, with in-house staff certified as lead auditors. AI governance & compliance →
U.S. and NATO registration is already in place, so teaming does not wait on paperwork. Every identifier is verifiable in a public registry.
Registry entries establish procurement eligibility. They do not imply endorsement by any government. Full detail is in our capability statement.
Yes, that is a common shape for our U.S. work. We hold the registrations a prime needs to add us to a team, and we are comfortable being the AI capability inside someone else's program rather than the name on the contract.
We are registered under Canada's Controlled Goods Program, and we scope export-control obligations explicitly during contracting rather than discovering them mid-delivery. Ask us for the export-control detail when you request the capability statement.
Wherever your requirements dictate. We deploy to AWS, Azure, GCP, NVIDIA DGX, your own datacentre, or fully air-gapped environments. Data residency is an architecture input, not an afterthought.
We can speak to CMMC; the level, scope, and status are supplied on request rather than published, so that what you get is precise and current. Tick the box on the capability statement request.
Tell us the program, the constraint, and the deadline. If we are not the right fit we will say so on the first call.