My fast-launch stack: shipping apps in days with .NET and Go

Launch fast or launch well? A lot of people assume that “moving fast” means reaching for the latest JavaScript framework. AI lets us ship software in very little time, but the technology you pick is still the foundation. Stability is part of it, but so is your ability to review and understand the code the AI writes for you.

AI tends to steer you toward React by default. My own productivity sweet spot, though, is somewhere many people dismiss as “too enterprise”: .NET and Angular.

An opinionated stack

Plenty of developers avoid .NET and Angular because they seem heavy. For me they’re what makes me faster. They ship with so much built in that I don’t spend time wiring up basic libraries or deciding how to lay out folders.

Compared with the dependency hell NPM can turn into, NuGet packages usually have very good support, which cuts down a lot on security issues. The stack also pushes me toward good practices: when the AI generates a feature, it lands in an architecture I already know and that scales easily.

Go where performance matters

Lately I’ve been adding Go for specific features. I use it when I need concurrency, since Go’s channels work really well for parallel processing, and when resources are tight: on my VPS I can run heavy processes in Go without wasting RAM, which is where other languages fall short.

PostgreSQL and its extensions

Every one of my projects uses the same database: PostgreSQL. With its extensions it stops being just a place to keep tables and covers a lot more:

• Timescale, for anything involving time series or advanced analytics.
• PGvector, for storing embeddings and running vector searches without leaving the main database.

My AI workflow: Antigravity and VSCode

For code generation I’ve switched to Gemini with Antigravity (I get a Google AI Pro subscription through my university). Its models and CLI have been a pleasant surprise: if a feature is well defined, it gets implemented in one or two iterations.

I haven’t dropped VSCode entirely, though. I tried Google’s extension for VSCode, but it tends to fail on complex prompts. There’s also a bigger reason: the Microsoft ecosystem. In IDEs like Antigravity I miss extensions Microsoft hasn’t released for third parties, such as the C# Dev Kit.

So my workflow is a hybrid:

1. I iterate and write most of the code in Antigravity with Gemini.
2. I run and debug in VSCode. With .NET Aspire, a single aspire run from VSCode gives me debugging ready to go with Microsoft’s native experience.

Orchestration with .NET Aspire

To tie everything together (.NET, Go and Postgres) I use Aspire. It lets me set up the infrastructure in a couple of lines of code and gives me full tracing and telemetry. To me, shipping a product isn’t only about the code working; it’s about knowing what’s going on inside from the first minute.

What I take from this

Going fast doesn’t have to cost you stability. With mature tools, Go where performance matters and Postgres covering more than you’d expect, I go from idea to deployment in days without losing track of what the AI is writing. That’s the part I care about most in the end: being able to read and understand the code I’m shipping.

Do you prefer a lightweight stack, or do you stick with tools that come with their own opinions?