Application Development Platform
Building new services used to take days, slowed by missing documentation and inconsistent environments. Matthew Westberg built the Application Development Platform, an AI-driven CLI that embeds Abnormal’s engineering patterns to scaffold Go and Python gRPC services, connect Postgres and S3 storage, and spin up Streamlit front ends.
November 4, 2025
NOTE: Demo visuals use either blurred real data or synthetic placeholders to protect customer privacy.
From Design Doc to Running App in Minutes
Matthew Westberg remembers trying to configure a single service and losing a whole day just finding the correct documentation. The archived materials, time zone gaps, and fragmented tribal knowledge turned a simple task into a slow relay across teams. That experience shaped the Application Development Platform: an AI-powered system designed to make service creation as fast and reliable as running a single command.
Building Without Delay
Before this platform, launching a new service meant manual setup, outdated guides, and inconsistent environments. Each new engineer had to learn the quirks of the stack by trial and error, often waiting overnight for answers from distributed teams.
The Application Development Platform solves this through an AI-driven CLI that scaffolds Go and Python gRPC services, provisions Postgres and S3 components, and builds Streamlit front ends directly from a technical design document. Deterministic guardrails ensure structure and compliance, while AI fills in the implementation details.
Matthew described it in his demo: “We can go from a tech design document to a functional app in about 10 minutes, which normally would take days or even weeks.” The CLI even standardizes how services are run locally, so engineers can start and test everything through a single "appdev run" command.

A Math Operations demo app showing Go and Python gRPC services, Postgres, S3, and Streamlit frontend.
Impact at Scale
The results are immediate. What once required weeks of setup now takes less than an hour. Engineers spend more time on product logic and less on configuration. Local and production environments behave the same, removing one of engineering’s most common failure points.
Service setup time reduced from weeks to under an hour
Reliable parity between local and deployed builds
Simplified onboarding for new engineers
Automated provisioning of infrastructure and boilerplate
The next phase will bring this capability to more teams, expanding the ability to go from idea to running application without waiting for context or configuration.
Early Reactions
Observers highlighted how the platform handled infrastructure and local runs without manual steps. They noted that what appeared in the demo environment matched production behavior, a shift that eliminates the usual debugging gaps between development and deployment.
This feedback shows a broader change in engineering culture. AI is no longer a side utility but a foundation for how software gets built. The Application Development Platform turns building services into an instant, repeatable process, accelerating both creativity and reliability across teams.
Problem
Engineers wasted time chasing tribal knowledge and resolving local-to-production mismatches, stretching setup out over weeks.
Solution
A CLI that enforces deterministic guardrails, generates full-stack services from specs, and runs everything consistently across environments.
Why it's Cool
Cuts setup from weeks to under an hour, streamlines onboarding, and lets engineers focus on innovation instead of infrastructure.
Technologies used:
- Golang
- Python
- Amazon S3
- Postgres
- Streamlit