AI Services
We run your
AI agents.
Start here
Audit
Free 1-week infra audit
Send us your current agent surface area. We map infra, latency, cost, and eval coverage and write you a 4-page audit. No charge, no obligation, no follow-up sales pressure.
- 4-page written audit
- Latency & cost breakdown
- Eval coverage map
- No obligation
Most teams pick this
Embed
4–8 week senior embed
Two senior AI engineers join your team. We ship the missing piece - a routing layer, a retrieval pipeline, an eval harness, a hardening sprint - to production, inside your infra, on your terms.
- 2 senior engineers, 4–8 weeks
- Weekly demo, monthly KPI
- Knowledge transfer baked in
- Source code transfer at handover
Post-launch
Run
Managed AI ops
We take the on-call. Model swaps, drift, schema drift, capacity, paging, monthly architecture review. 99.9% SLA with a real runbook behind it. You ship the product; we keep it alive.
- 99.9% uptime SLA
- On-call rotation (HKT hours)
- Cost-cap on $/1k tokens
- Quarterly model-swap review
The senior side of the practice
Strategy, on tap.
Where you need it.
Every engagement starts with a discovery brief - a written document the whole team agrees with, not a slide deck. The same engineers who run the system help you decide what to build.
Strategy
AI roadmaps, model selection, build-vs-buy, vendor review.
- Roadmap
- Vendor review
- Build vs buy
- Risk model
Implementation
Embed as senior engineers - ship the first system with you.
- Agent loops
- RAG pipelines
- Evals
- Cost engineering
Operations
Own the system post-launch - keep it fast, cheap, correct.
- On-call
- Model swap
- Drift watch
- Cost cap
Common questions
On engagements.
Which shape is right for us?
Can we change shapes mid-engagement?
What does 'managed' actually cover?
Do you sign MSAs, DPAs, security questionnaires?
Is the audit really free?
What if we're not based in Hong Kong?
Start with the audit