Local AI consulting

Self-host AI on your own server — no cloud lock-in, no data leaks, no API bills that scale with your user count.

The KoishiAI site you are reading runs on the same stack we deploy for clients — real pipeline, real models, real hardware, not a slide-deck demo. Engagements are handled by the editor directly; no sales agency, no junior handover.

Service packages

Local AI Setup

Install and run LLMs on your own server

For: Hospitals, clinics, law firms, accounting advisors, small banks — anyone needing PDPA compliance without shipping data to cloud providers.

✓ Workload + hardware assessment (throughput, VRAM, concurrency)
✓ Install Ollama / vLLM + Open WebUI on your server
✓ Model selection and loading (Qwen3, Gemma, Llama)
✓ Backup, logging, and monitoring setup
✓ Hand-off, documentation, and 30-day support

From THB 30,000

Scope-dependent pricing. On-site in Bangkok metro included; outside-metro travel billed separately.

Email to inquire

Thai AI Content Pipeline

Editorially governed AI content, not generic SEO slop

For: Thai startups / SMEs / brands needing systematic Thai-language thought leadership, not one-off freelance articles.

✓ 15-20 SEO-ready articles per month (native Thai + English)
✓ Multi-source fact-checking with inline citations
✓ TL;DR + Key Facts + FAQ on every piece (AI-search ready)
✓ Hero image + alt text + schema.org markup
✓ Live admin dashboard for pipeline status

From THB 30,000 / month

3-month minimum retainer. Topic scope and editorial policy agreed up front.

Email to inquire

Local AI Workshop

Teach your team to run LLMs locally in 2 days

For: Internal IT teams or dev groups wanting hands-on experience, not another slide deck.

✓ 2 full days hands-on (8 hours per day)
✓ Install Ollama + WebUI on each participant's machine
✓ Run Qwen3, Gemma, Llama and compare throughput
✓ Build a RAG + custom chatbot for your team
✓ Q&A session, slide deck, and starter-code repo

THB 15,000 / person (min 4) or THB 60,000 / day in-house (up to 10)

VAT added only when invoiced to a corporate entity.

Email to book

How engagements work

1

Reach out — free 30-min call

Email me with a rough scope; I reply within one business day and we set up a free video call to see whether I can actually help.
2

Proposal + clear scope

After the call, I send a PDF proposal with deliverables, timeline, price, and terms. No hidden fees, no pay-for-proposal.
3

50% deposit — work begins

Once approved, 50% deposit via PromptPay or bank transfer. Weekly updates from day one.
4

Delivery + support

Deliverables handed over after final payment. 30-day support covers bug fixes for what was delivered.

Why an independent researcher, not an agency

● No sales layer, no junior handover — you speak directly with the person doing the work
● Reachable pricing (from 30k vs. agency retainers 150-300k/month)
● Public proof of work — KoishiAI + GitHub + benchmark articles audit-able before hiring
● Deep specialist in one niche (local LLM on consumer GPU), not a generalist consultant
● No vendor lock-in — we install on your hardware, you own it

What we do not take on

✗ Writing promotional articles on KoishiAI for a client — violates editorial integrity
✗ SEO spam, keyword stuffing, or AI SEO junk content
✗ Cloud-dependent AI setups for PDPA-sensitive work (we are on-premise specialists)
✗ Outcome guarantees tied to external platforms (Google ranking, FB engagement, etc.)
✗ Anything contradicting our editorial standards

KoishiAI editorial independence

The services above are the editor's personal business — separately from KoishiAI (the news site). We do not publish articles on KoishiAI to benefit clients, do not add affiliate links, and do not alter editorial content for money. Full details at our editorial standards page.

Contact

Email kiatdamrong@gmail.com with a brief scope of what you want to accomplish. I reply within one business day.

Email: kiatdamrong@gmail.com