We build and own digital products on a shared AI foundation, with one accountable lead each. Owned products remain the core. Two invitation-only client website pilots give the team supervised real-world practice; they do not turn the lab into an open agency.
One studio, seven intended product tracks, and two clearly separate client-pilot slots. Four tracks still need a written ownership decision, so read the boundaries literally.
Chnai Lab is one parent brand that builds and keeps many small digital products. Every product has one lead — a single builder who owns it end to end and holds a real stake in what it becomes. No committees, no handoffs to people who never touched the code.
Alongside the portfolio sits a shared AI/ML capability — models, pipelines, and tooling we verify once and reuse where the product need and safety boundary fit. Reuse can reduce later build cost; it is never mandatory.
Our current job is to verify ownership and define a real finish line for every intended product track, in sequence. Alongside that completion lane, we may run up to two bounded client pilots once each partner explicitly accepts a narrow brief. Client work never silently jumps ahead of the verified lab-owned portfolio.
Not a fantasy deck — a controlled path from repository inventory to verified finish lines and consented proof.
Verify ownership and current truth for every product. Give each one a finish line, accountable lead, next action, and evidence requirement.
Run one owned-product completion lane at a time. Closest-to-proof work moves first; paused and research tracks stay honestly labeled.
Use two simple websites as real delivery practice. Scope, content, launch, credit, support, and handoff stay under written client approval.
Publish only shipped evidence, accurate contributor roles, and client-approved outcomes. Harder work enters only after current commitments close.
Six rules we actually behave by — not wall posters.
We show the working thing first and talk about it second. No announcing what doesn't exist yet.
Every product has a single owner who builds it and holds a genuine stake — not a manager, the maker.
If we solve it well once, it goes into the shared layer so the next product gets it for free.
One owned-product completion lane plus up to two bounded pilots. A third client project is out of bounds.
Product stakes, pilot scope, costs, revisions, support, ownership, renewal, termination, handoff, and public proof are written before they become promises.
We'd rather tell you the real stage than win the meeting. Stage 0 said plainly beats a fake Stage 3.
The same path every product walks, with one lead carrying it the whole way.
A real problem we keep running into. We dig until we know it's worth building and what the shared layer already covers.
The smallest version that proves the idea — rough, fast, and standing on the existing AI foundation.
Heads-down work to a usable product. Security and the shared models come in here, not bolted on later.
Into real hands. We watch real usage, fix what's wrong, and find the first people who'll pay or partner.
If a product outgrows the lab, we help it become its own thing — and keep a real stake in what we built.
A small crew of student-engineers in Phnom Penh. Everyone builds; everyone owns something.
Leads Chomkar and sets the bar for what the lab ships.
Leads SAT Digital and keeps security & trust solid across every product.
Leads the shared AI/ML capability and verifies where reuse is justified.
Handles partnerships, buyers, grants and getting products in front of real users.