Engineering

Engineering and advertising are two practices with one billing model: a flat monthly retainer, a small accountable team, and compute billed at the public rental rate. No templates, no percentage of spend, no subcontracted creative, no hourly change-order theatre.

Discovery is two weeks against your actual systems - APIs, warehouses, ad accounts, order backends - and ends with a scoped build plan and a number. If we are the wrong shop, that is the outcome. If we are not, Core is $14,000 a month (~120 senior hours), Extended $32,000 (a three-person pod), Embedded from $68,000 (operates as your department).

Engineering takes on software you cannot buy off the shelf: internal tooling, data pipelines, pricing engines, inference services, legacy migrations. Where the workload needs GPUs we run it on our fleet, which is usually why the unit economics work at all. Source, infra and runbooks are yours from day one. Deploy on Sundancæ or your cloud, written up front.

Advertising reads ad platforms, analytics and orders through APIs, trains a spend model on contribution margin, and buys media without taking a cut of spend. Creative stays in-house. Holdout tests validate incremental revenue. At $150k monthly spend, $14k is a 9.3% effective fee and falls as spend grows - the opposite incentive of a 12–15% shop.

Minimum term is three months, then thirty days’ notice. Compute used on our hardware still hits the rate card. If the job runs on your cloud, that line is zero. The client portal is where connected Meta and Google accounts show up once credentials are on file - that is a product of the advertising practice, not a separate SaaS.