Architect.
Builder.
Founder.
Software Solution Architect · iA Financial Group · Founder, Ghasi Technologies
● AI-first & offline-first by design — foundational, not bolted on.
I'm Asmar Momand. I take ideas from whiteboard to production — architecture, solution design, cloud infrastructure, and the code that ships. Solo when the team is small. With a team when it's not. The architecture and the implementation are the same conversation; I refuse to let them be different jobs.
That conviction comes from a specific kind of practice. I am close enough to production that the constraints I prescribe are constraints I am still living with. Close enough to the team that I review the PR that implements the ADR. Close enough to the on-call rotation that an architectural shortcut I sign off on is a page I'll personally answer at 02:00.
I started in 2012 at the World Bank, writing software where mistakes cost institutional credibility, not just sprint velocity. Three years at GIZ in Afghanistan taught me what public financial management software actually has to do — keep working when everything around it is improvising. After moving to Canada, I spent two years owning the front-end and parts of the cloud stack of a private RoboAdvisory platform, and the last five at iA Financial Group — working through senior developer, tech lead, and senior solution designer to my current role as software solution architect, designing for private FinTech, wealth management, and insurance lines of business.
In parallel, I've been founding Ghasi Technologies — a portfolio of enterprise platforms across health, education, hospitality, government intake, and the shared infrastructure underneath them. It's how I keep my hands on the keyboard. I believe architects who don't ship lose calibration.
Every platform in that portfolio is AI-first and offline-first by design — both treated as foundational architecture, never as features bolted onto a connected, non-AI core. A single AI Gateway is the only egress for any LLM call; every artifact carries provenance metadata; budget and safety are enforced at the port, not at the call site. The client is the primary actor; reads and writes go through a local outbox; vector-clock sync reconciles when the network returns. I have working experience with the patterns that make both of those true — multi-provider LLM gateways, RAG, on-device inference for offline clients, AI evaluation pipelines, per-aggregate conflict resolution, signed offline bundles, and the offline-sync protocols underneath.
I work in English, Pashto, and Persian. I mentor students through Harvard's CS50. I think about distributed systems the way some people think about chess: a small number of patterns, an enormous space of consequences.
I'm currently open to senior, staff, and principal architecture roles, to independent consulting (advisory through embedded), and to founder-shaped problems where one architect-builder in the first ninety days saves you a team in the next twelve months.
Cloud experience
GCP · AWS · Azure
Architecture, infrastructure, AI services. Not slideware.
Certifications in progress: AZ-305, AI-102, AWS SAA-C03.
Languages
- English (full professional)
- پښتو (native)
- فارسی (native)
Based in
Toronto / Ontario / Remote-first across EU & Middle East
What you actually get
Not a slide deck. Not a 90-page Word document nobody opens twice. Concrete artifacts that move a decision or unblock a team within the week.
Context maps
Bounded contexts, relationships, and the integration patterns between them — annotated.
ADRs
Each names the alternatives, the tradeoffs, and the conditions under which I'd revisit it.
Service blueprints
API contracts, event schemas, data models — the shape the team will build to.
Risk register
What can go wrong, what it costs if it does, and what we'd do about it.
POVs & POCs
When the answer is genuinely unknown, evidence beats opinion.
Production code
When the team is lean, I write the part that ships. With CI gates, on-call, the boring parts that matter.