Services
- Sheet
- A-010
- Set
- Practice
- Status
- Issued
Three ways we work on your software.
Every engagement is led by a senior engineer, scoped and priced before it starts, and delivered through the same reviewed engineering system. What changes is where your software is today.
A-010PracticeIssued
-
A-050
AI product engineering
Build AI into the product.
LLM features, AI-native products and multi-model systems, engineered to run in production with evaluation, cost control and human review designed in.
For: Product teams adding AI, and companies building an AI-native product.
AI product engineering → -
A-100
Engineering
Build, integrate, modernize, harden.
Full-stack product engineering, platform and backend work, integrations, modernization and release engineering, delivered by engineers directing AI-assisted work.
For: Teams that need real engineering capacity, not a headcount number.
Engineering → -
A-200
Prototype to Production
The prototype proved the idea. Now build the product.
An assessment of an app built with Lovable, Bolt, Replit, Cursor or Claude Code: what a stranger with dev tools could reach, what fails under load, what to keep, and a priced plan to fix it.
For: Founders and teams whose AI-built app is about to meet real users.
Prototype to Production →
Domains and disciplines
Where we have built, and what we do.
Depth is stated per domain, not claimed across the board. Each discipline links to something you can check.
Domains
- Healthcare In production A multi-tenant telehealth platform, from intake to e-prescribing, live and in use.
- SaaS, AI products & complex systems In production Multi-tenant SaaS with database-enforced isolation, and the AI engineering system we run ourselves.
- Legal Built, not in paid use AI-assisted casework tooling for criminal defense, every output gated behind attorney review.
- Real estate Prototype An investment-analysis platform that scores opportunities and backtests its own projections.
Disciplines
- Full-stack development Web applications, APIs and internal tools, built end to end. 940+ merged pull requests →
- AI and LLM engineering Model routing, evaluation, cost limits and human review, in production. Our own multi-model engineering system →
- Security engineering Security reviews and hands-on testing of web apps and APIs, access control and authentication, cloud and infrastructure hardening, secrets and dependencies. We do not sell compliance certifications. Field note: row-level security first →
- Data and databases Schema design, migrations and Postgres that stays correct as the product changes. 2,500+ migrations in one platform →
- CI and release engineering Pipelines, deploy gates and environments where nothing ships unchecked. 16,800+ CI runs →
- Testing and QA Automated tests written with the code, and failing tests written before fixes. 20,900+ automated tests →
- Integrations Third-party APIs, webhooks and background jobs, with the failure cases handled. 370+ server functions in production →
- Modernization Taking over existing code: test it first, then change it, then replace what has to go. A rewrite is a decision, not a default →
Not sure which fits?
Tell us where the software is today. An engineer will tell you which of these it needs, or whether it needs us at all.
Next sheet · A‑050