Questions
The questions buyers actually ask, answered plainly.
No hedging and no sales language. If an answer here rules us out for your situation, that is the answer doing its job.
Engagement
Engagement questions.
What does a first engagement usually look like?
A possible first step is a non-binding scope discussion. Any USD proposal follows understanding your needs; deliverables, timing and terms require agreement. An enquiry does not book work.
How small is too small?
If the work is a single feature on an existing healthy codebase, you are better served by a contract engineer than by an architecture practice. We will say so rather than take the work.
Do you work with pre-revenue founders?
Pre-revenue founders may discuss architecture and sequencing. If there is mutual fit, an equity or fee-based partnership may be considered under separately agreed delivery and operating terms; no previous partnership frequency or availability is claimed.
Can you work with our existing development team?
We can work alongside your existing team, with responsibilities, authority and staffing agreed in scope.
Delivery
Delivery questions.
How long until something is in production?
The schedule depends on scope, access, integrations and review. A phase may produce a validated prototype or a release increment; production requires agreed acceptance and authorization.
What technologies do you build on?
Typescript and React on the front, Postgres-backed services behind, with AI surfaces built against model gateways rather than a single vendor. Stack choices follow the constraint, not a preference — that reasoning is recorded in the decision log.
Who owns the code and the IP?
Ownership of custom deliverables, repositories, reusable methods and third-party components is set in the signed agreement. Open-source and licensed dependencies retain their applicable terms; handover and export rights are agreed before work begins.
What happens after handover?
Handover artifacts, operating support, accountable contacts and any support window are agreed in the signed scope. No ongoing operation is implied by an inquiry.
AI and compliance
AI and compliance questions.
Can AI features be deployed inside a regulated boundary?
Yes, with the boundary designed first: defined enclaves, recorded human approval as a control, evidence generated as a by-product of the workflow rather than assembled before an audit. Applicable requirements need scoped review; this page does not establish certification or delivery history under those frameworks.
How autonomous should our agents be?
Start assisted, earn enabled, and only reach agent-run where the authority boundary and evaluation harness exist. Undefined authority is a failure mode we assess.
Will you sign our DPA, MSA and security questionnaire?
We can review proposed agreements and security questions during scoping. Signing, insurance evidence, access, commercial model and operating responsibilities require confirmation.
Do you use our data to train models?
Data use and model-provider processing must be agreed for the specific engagement before any data is shared. Do not submit confidential material in initial intake.
Still deciding
Bring the question that isn't listed here.
Describe the decision and constraints. Scope, availability and any follow-up require confirmation.