AI Consulting & Development · Las Vegas + Remote

AI consulting & development — without the agency, the handoffs, or the per-token bill.

Thirty years shipping software, now focused on AI. I architect, build, and host custom models, automation, and integrations myself — on infrastructure I own, so your data never leaves my hands and there's no meter running on every request.

Send me an email

The fastest way in — tell me what you're building, and we'll set up a call from there.

Building for the web since 1994 Self-hosted AI infrastructure One senior engineer, start to finish No account managers

The problem with most "AI help"

You don't need another AI strategy deck. You need it built.

✕ 01

An agency scopes it, then hands you to a junior who's learning on your budget — you never talk to whoever actually writes the code.

✕ 02

The "AI solution" turns out to be a thin wrapper on someone else's API, and the bill grows every single time you use it.

✕ 03

To make it work, your proprietary data gets piped off to a third party's servers — with fine print you don't control.

✕ 04

Six weeks of meetings and discovery calls later, there's still no working software you can actually put in front of a customer.

What I do

Consulting when you need direction. Development when you need it built. Both from one person.

Phil Blancett

FOUNDER · OAKWEB.AI · LAS VEGAS

I'm Phil — I've been building software for thirty years, and I do this myself. The person you scope the problem with is the person who architects it, writes it, ships it, and keeps it running. No translation layer, no "let me check with the dev team," no surprise handoff. Whether you need an hour of straight answers or a full system stood up, you're talking to the builder.

Why work with me

Senior engineering, minus the agency tax.

Talk to the builder

The person who answers the phone is the person writing the code. No account managers, no handoffs, no telephone game.

Own your stack

Models can run on my hardware — no per-token API meter that punishes you the more your product succeeds.

Your data stays private

Inference on infrastructure I control. Your proprietary data doesn't get shipped to a third party's servers to work.

30 years of shipping

From a garage in Lake Tahoe to a thousand servers in San Francisco to an AI studio in Vegas. This isn't my first rodeo.

Working software, not slides

You get a system you can put in front of customers — not a strategy deck and an invoice for the discovery phase.

Frontier when it fits

Want GPT or Claude for the hardest tasks? Fine. I build hybrid — local for privacy and cost, frontier APIs where they earn it.

The differentiator

I don't rent AI capacity. I own it.

Real infrastructure — not a login to someone else's API

Your models, on my metal.

Most "AI companies" are a login to somebody else's API. I run a real cluster of my own — high-memory machines capable of hosting large open models end-to-end. That changes what I can offer you:

Private by default. Sensitive data can be processed entirely on hardware I control, never leaving for a third-party API.

No usage meter. Self-hosted inference means your costs don't balloon with every request as you scale.

512GB

Unified-memory nodes — enough to host the big open models (DeepSeek, Qwen) in-house.

GPU

Dedicated accelerators for image, video, and low-latency generation.

RAG

Private retrieval + embeddings so a model can reason over your documents — on your terms.

MACHINES ON-PREM

~2.2 TB

UNIFIED MEMORY · AI CLUSTER

671B

PARAM MODEL, RUN LOCALLY

PER-TOKEN API COST

4×

Mac Studio

M3 ULTRA · UP TO 512 GB

The EXO cluster — Thunderbolt-linked 512 GB nodes that host the largest open models end-to-end.

3×

DGX Spark

NVIDIA GB10 · 128 GB

Grace-Blackwell GPUs for image & video generation and CUDA workloads.

1×

Minisforum MS-S1

RYZEN AI MAX · 128 GB

Always-on inference, a private RAG archive, and document embeddings.

5×

Mac Mini

APPLE SILICON · M-SERIES

Edge nodes running automation and agents around the clock.

Custom AI Systems AI Automation Chatbots & Agents RAG / Document AI API Integrations Web Apps Websites Backend Systems

The honest math

"Does AI cost money?"

Yeah — a lot. The only question is where it lands. Here's my real AI stack, every month, on autopay:

OAKWEB.AI · MONTHLY AI STACKSINCE 1994

ChatGPT Pro · frontier model$200.00

Claude · frontier model$200.00

OpenRouter · model API credits~$150.00

Higgsfield · video & image gen$49.00

Perplexity · research$40.00

Lovable · app builder$25.00

Firecrawl · web scraping$19.00

Apify · scraping$0.00

MONTHLY TOTAL≈ $683 / mo

…all of it running on ~$53,400 of AI hardware sitting in my office.4× MAC STUDIO · 3× DGX SPARK · 1× MINISFORUM · 5× MAC MINI

You pay $200.

I cover all of this.

That's the whole point of hiring someone who already owns the stack. No subscriptions to juggle, no surprise per-token bills, no $53k hardware run to bankroll. One flat rate — the AI overhead is my problem, not yours.

Send me an email

How it works

From "here's the problem" to shipped.

Free consult

Send me an email. Tell me what you're trying to do. You'll get straight answers, not a sales funnel.

Scope & architecture

I map what it takes — what runs local, what runs frontier, what it costs, how long.

Build

I write it. You see real progress, not status meetings. Working software, fast.

Run & maintain

I can host and keep it running on my own infrastructure — or hand it off cleanly. Your call.

Straight answers

Common questions.

Do you actually build, or just advise?

Both, and that's the point. I take consulting-only engagements when you just need direction — and I build and ship full systems when you need it done. Same person either way.

Can you still use OpenAI or Claude if I want frontier models?

Absolutely. I build hybrid systems: local open models for anything privacy-sensitive or high-volume, and frontier APIs like GPT or Claude for the hardest reasoning. You get the right tool per task instead of one vendor's bill for everything.

Is my data actually private?

When it needs to be, yes. I can run inference on hardware I own, so your proprietary or regulated data never has to leave for a third-party API. We decide together what runs where.

What does it cost?

It depends on scope — a consult is free, and I'll give you a real number before any work starts. No discovery-phase invoices. See packages & pricing for a sense of range.

Do you only work with Las Vegas businesses?

I'm Vegas-based and love local work, but most of what I build is remote-friendly. Same senior engineer whether you're across town or across the country.

How fast can you start?

Because it's just me, I take on a limited number of builds at a time — which keeps quality high and timelines honest. Reach out and I'll tell you exactly where things stand.