The Velocity Framework

The Velocity Framework is Limestone’s end-to-end methodology for AI-augmented software delivery. It defines how we structure teams, how engineers collaborate with AI day to day, and which quality and governance rules keep the whole system safe.

This section is written for people who need to use the framework in real projects, not for marketing. It is deliberately practical, opinionated, and technical.

Why Velocity

Modern teams sit between two broken extremes:

• Traditional outsourcing / staff augmentation: scale by adding people, not by improving workflows. Knowledge is fragile, onboarding is slow, and quality depends too much on individual heroes.
• “Just add AI” tooling: a few plugins sprinkled on top of old processes. Developers get autocomplete, but planning, review, and delivery rhythms stay the same.

Velocity takes a different stance:

• AI-native workflows: we redesign planning, coding, testing, and documentation around AI from the start.
• Pods instead of loose squads: small, focused teams with clear ownership and shared rules, not ad-hoc collections of contractors.
• Measured outcomes: we care about lead time, failure rates, and rework—not just how many tickets moved to "Done".

Internal projects have already exercised this workflow at scale. The framework here is the distilled version of those lessons, not a theoretical process.

Who This Is For

The primary audiences are:

• Limestone engineers working in Velocity Pods.
• External engineers considering joining a pod and wanting to understand how we actually work.
• Technical client stakeholders (architects, tech leads, CTOs) who need to understand our delivery model and constraints.

The tone is developer-first and execution-oriented. We avoid commercial packaging here; pricing, sales language, and case studies live elsewhere.

How To Use This Framework

Use this section as a map of how work gets done:

• Core Philosophy explains the mindset behind Velocity: the manifesto, the AI-augmented engineer role, and how product functions plug into the same model.
• Governance, Quality & Metrics documents the non-negotiable rules: quality gates, security/privacy expectations, and how we measure flow.
• AI Delivery Stack describes the capabilities we expect from AI-native IDEs, context integrations, and model choices (without tying the framework to a single vendor).
• AI-Augmented Engineering Workflow formalizes the 6-step loop we expect engineers to run on every significant change.
• Product & Delivery Workflows explains how pods structure work, how product and delivery collaborate with AI, and how we communicate progress.
• Reference Library collects reusable patterns, templates, and integration catalogs that instantiate the framework.
• Appendix contains glossary, FAQ, and pointers to contribution guidelines.

The /onboarding area builds on top of this framework with a guided path for new engineers. It does not redefine the methodology; it references the pages here and adds phased training, environment setup, and internal-only details where needed.