/how-we-work

How we work.

The boring-but-important questions, answered. Engagement shapes, lifecycle, security posture, AI policy, pricing.

Engagement models

Four shapes. Pick the one that fits the work.

[ 04 / 04 ]
01 · fixed-scope
Fixed-scope project

Clear specs, fixed price. Best when you know what you want and want a guaranteed quote.

typical: 6–24 weeks
02 · dedicated-team
Dedicated team

Evolving scope, monthly rate. Best when scope will change as you learn.

typical: 3–12 months
03 · staff-aug
Staff augmentation

Senior hands into your team. Best when you have a tech lead and need experienced engineers.

open-ended
04 · discovery
Discovery & advisory

Pre-build clarity, architecture review, second opinion. Best before committing to build.

typical: 1–3 weeks

The AI-accelerated lifecycle

Where AI shows up. Where it doesn't.

[ 07 stages ]
// AI
// Human
01
Discovery & spec
Requirement analysis, gap-finding, diagram generation from descriptions.
Scoping decisions, prioritization, stakeholder alignment.
02
Architecture
Decomposition, tech comparison, estimation sanity-checks.
Architecture choices, trade-offs, integration strategy.
03
Implementation
Scaffolding, integration code, repetitive logic, test scaffolds, type definitions.
Business logic, edge cases, performance-critical paths.
04
Testing
Test generation, edge-case discovery, fuzzing inputs.
Test strategy, acceptance criteria, exploratory testing.
05
Review
First-pass review, style and security pattern detection.
Review for intent, architecture compliance, business correctness.
06
Documentation
Drafts from code, diagram generation, keeping docs current.
Explaining why, not just what.
07
Operations
Log analysis, anomaly detection, runbook drafting.
Production incidents, root cause, customer communication.

Tech stack

Pragmatic, not religious.

languages
Python · TypeScript · .NET · Swift · Kotlin
frontend
React · React Native
backend
Node · FastAPI · Django · .NET
data
PostgreSQL · MongoDB · ClickHouse
ml / ai
PyTorch · scikit-learn · Hugging Face · OpenAI / Anthropic / Google APIs
infra
AWS · Azure · GCP · Docker · Kubernetes

Security, data, AI policy

The part vendors lose deals on.

01 · code & ip
Yours from day one.

All code and deliverables transfer with the project. NDAs default. OSS contributions only with explicit permission.

02 · data
GDPR by design.

Minimum data needed. Production isolated from development. Destroyed at project end unless contractually retained.

03 · ai-policy
Tools documented per project.

Zero-data-retention enterprise plans for any project touching client data. Never used to train shared models. Approved tool list on request.

04 · security
Pen-test friendly.

Vault-based secrets. Dependency vulnerability scanning in CI. Audit trails default.

Quality

What good looks like, to us.

Senior engineers on every project. Two-week iterations with working software at the end of each. Continuous deployment where appropriate. Test coverage non-negotiable. Honest about delays. Honest about scope.

Where we say no

The work we don't take on.

Pure design (we partner with design teams). "Build me the next [unicorn]" briefs. Speculative crypto, gambling, surveillance, anything our team wouldn't be proud of. Outcomes we can't deliver — even when it costs us the project.

Indicative pricing

Roughly what to expect.

EUR · indicative · 2026
01 · discovery
Discovery
€[range]
1–3 weeks
02 · mvp-build
MVP build
€[range]
8–16 weeks
03 · team
Dedicated team
€[range] / month
3–5 engineers
04 · staff-aug
Staff aug
€[range] / month
per engineer

Get in touch

Questions we didn't answer?

Book a 30-minute scoping call. We'll get specific about your project.

Book a scoping call See case studies →