Forem: David Smith

We’re Opening Our Google Cloud Infrastructure to Researchers, Healthcare Teams & Builders

David Smith — Fri, 02 Jan 2026 03:25:48 +0000

We’re Opening Our Google Cloud Infrastructure to Researchers, Healthcare Teams & Builders

Most people don’t struggle with writing code.

They struggle with cloud setup, billing, IAM, scaling, and operational overhead.

At David Labs, we’ve been running production workloads on Google Cloud Platform for our own products (AI, healthcare, data platforms). We’re now opening that same infrastructure to others who want to explore, research, or build on GCP without managing everything themselves.

Who this is for

AI research fellows experimenting with Google’s AI models
Healthcare teams exploring analytics, HL7 pipelines, or secure data processing
Developers & founders validating SaaS ideas
Teams learning GCP hands-on without billing anxiety

What we provide

GitHub Enterprise
- Org, repos, CI/CD pipelines
Google Cloud Platform
- Cloud Run, IAM, OAuth
- Any GCP service (including AI & data services)
MongoDB Atlas
- Premium M10 cluster
Stripe
- Billing integration (if you plan to monetize)
Hubble
- Our in-house data visualization tool
- https://hubble.davidlabs.ca
Subdomain hosting (yourapp.davidlabs.ca)

You focus on code, research, or product logic.

We handle infra, scaling, security, and cloud plumbing.

Engagement models

Revenue share: 70% you / 30% us (Stripe auto-split)
Monthly rent: usage-based + 15% infra management fee
Research / AI workloads: custom, project-based setup

✔ No equity

✔ No upfront cloud bills

Why we’re sharing this

This is an experiment. We’re validating whether managed, production-grade GCP access is valuable for researchers and builders who don’t want to reinvent infrastructure.

If you’re exploring GCP and want a managed environment, I’d love feedback.

📩 infrastructure@davidlabs.ca

🌐 https://www.davidlabs.ca

Tags: googlecloud ai healthcare saas

How to build an app in 2025.

David Smith — Wed, 10 Dec 2025 23:44:01 +0000

Let’s start simple.

It’s 2025. AI is everywhere. But building a useful app still comes down to one thing:

Solve your own problem first.

Example: Building a simple DB explorer

Let’s say you want to build a simple DB explorer (like hubble.davidlabs.ca).

What do you actually want from the app?

Explore MongoDB data
No query syntax
No learning aggregation pipelines
Just ask questions in natural language

That's it. That’s the entire spec.

This is how you build apps in 2025:

Start with the most annoying part of your workflow
Remove the friction
Let AI fill the gaps

Not “ship an MVP.”

Not “validate an idea.”

Solve your own pain so deeply that other developers say:

“Oh thank god, I needed this too.”

1. The “Why” is more important than the “What”

Before writing a single line of code, ask yourself:

Why does this app deserve to exist?

For me: MongoDB is great, but exploring it was painful.

Aggregation pipelines, Compass tabs, Atlas views… too many steps.

My “why” became:

“I want clarity without syntax.”

Once the why is clear, the what becomes obvious.

2. AI is your co-developer in 2025

In 2025, building apps looks like this:

Ask GPT to scaffold your backend
Use AI to design your UI
Let AI generate validation & types
Let AI write Dockerfiles
Let AI write text, onboarding, docs
Let AI generate fake data
Let AI refactor messy code
Let AI help debug faster

AI isn’t replacing developers.

AI is accelerating developers.

3. Add intelligence only where it matters

Most apps don’t need “full AI.”

They need tiny intelligent features that remove friction:

Natural language input
Smart suggestions
Autocomplete
Anomaly warnings
Auto-configuration
AI-assisted logs
One-click insights

Hubble works because intelligence is placed exactly where the frustration was:

querying MongoDB.

90% value, 10% code.

4. Ship early. Ship simple. Ship real.

Your v1 doesn’t need:

fancy design
animations
a logo
20 features
full documentation

Your v1 needs one thing:

It should work.

Everything else is optional.

5. One unfair advantage beats 100 features

You don’t need a huge roadmap.

You need one thing that makes people go:

“Hold up… this actually saves time.”

For Hubble, that one thing was:

Ask MongoDB questions in English → see the generated pipeline.

That’s the unfair advantage.

Your app needs one too.

Final thought

The way to build apps in 2025 isn’t about frameworks, stacks, or patterns.

It’s about:

reducing cognitive load
using AI to accelerate development
solving a problem you feel deeply
creating clarity
shipping early
iterating in public

If you want to build something meaningful this year:

Start with something that annoys you.

Solve it deeply.

Ship it.

Everything else will follow.

Meet Hubble: A Natural-Language Observatory for MongoDB

David Smith — Tue, 09 Dec 2025 23:11:29 +0000

🚀 Introducing Hubble — An NLP-Powered MongoDB Explorer (Beta)

For years, MongoDB has been one of my go-to databases — flexible, fast, and great for modern applications.
But exploring the actual data inside it?
That was always harder than it needed to be.

A simple question like:
“How many users signed up today?”

always turned into a mini quest through aggregation syntax, roles, connection strings, and multiple tabs in Compass or Atlas.

Eventually I got tired and wrote a small Python script for myself — something that just showed me the answers without the friction.

That tiny script kept growing.
Then I added natural language.
Then visualizations.
Then anomaly detection.
And suddenly I realized:
“Maybe other developers want this too.”
That's how Hubble was born.

🌌 What is Hubble?
Hubble is a Database Observatory for MongoDB.
It lets you explore your database using natural language instead of queries.
Think of it as:
“What if MongoDB had a conversational UI?”
You can ask Hubble:
“Show me users from Canada”
“Find failed transactions in the last 24 hours”
“Group orders by month”
“Detect anomalies in payments”
…and Hubble generates the underlying MongoDB query + visualizes the results.
No syntax.
No mental overhead.
Just answers.

🧠 Why I built it
Not because I wanted to build a startup.
Not because I wanted to build an AI tool.
I built it because I simply wanted clarity.
Databases hold the actual story of how your system behaves — but reaching those insights often requires digging through layers of complexity.
Hubble removes that layer.

🛠️ How to Try the Beta (Docker)
Hubble is launching as a Docker-first developer preview so it's easy to try locally.

docker stop hubble; docker rm hubble; docker pull davidos366/hubble:latest && docker run -d -p 3366:3366 --name hubble davidos366/hubble:latest

Connect your MongoDB URI → explore your data.
That's it.

🔭 What it can do today

Natural language → MongoDB queries
Document visualization
Schema exploration
Basic anomaly detection
Query transparency (you see the generated pipeline)
Fully local (no cloud required for the Beta)

🎯 What’s next

I’m currently working on:
A hosted SaaS dashboard
Saved queries
Team access
Telemetry + insights
Richer charts
Support for more databases (Postgres next?)

🙌 I’d love your feedback
This beta is intentionally early — I want to learn from real developers before shaping the full release.
If you try Hubble and have thoughts (good or bad), I’d genuinely love to hear them.

Thanks for reading — and for checking out Hubble.
— David

We submitted the same app in the bolt Hackathon!

David Smith — Fri, 25 Jul 2025 17:04:20 +0000

David Smith

Jul 25 '25

Our First Product! - ZUBO

#programming #webdev #javascript #react

Comments

1 min read

Our First Product! - ZUBO

David Smith — Fri, 25 Jul 2025 17:03:22 +0000

Hello Everyone,

We want to present to the world - ZUBO

An Intelligence game powered by a Native AI developed at David Labs.

Please give it a try and let us know how it goes.