AI Engineering
Forward Deployed AI Engineering
You have won (or are chasing) a Korean enterprise customer and have no engineer in-country. Rent one. Kuro Tech embeds with your customer to deploy, integrate and localise your AI product — then hands over cleanly.
What a forward-deployed engineer does here
- Deploys LLM applications into Korean customer environments: Claude, OpenAI, AWS Bedrock, Azure, on-prem or VPC
- Builds retrieval (RAG) over Korean documents, with evaluation in Korean
- Designs agent workflows with human escalation and audit dashboards
- Integrates with customer data: SQL, document stores, internal APIs, messaging (KakaoTalk, Slack, Teams)
- Handles Korean language, data residency and compliance questions with the customer's IT and security teams
- MLOps on AWS; computer-vision integration where needed
- Trains the customer's team and writes the hand-over documentation
Engagement models
| Model | Korean client (KRW, ex-VAT) | International client (USD) |
|---|---|---|
| Day rate | ₩700,000–900,000 | $550–700 |
| Embedded, 2 days / week (monthly) | ₩5,500,000 | $4,500 |
| Embedded, 3 days / week (monthly) | ₩8,000,000 | $6,500 |
| Fixed-scope pilot (2–4 weeks) | ₩8,000,000–15,000,000 | $6,000–12,000 |
| Discovery & architecture workshop (1 day + written plan) | ₩900,000 | $700 |
Swipe sideways to see the whole table.
Minimum engagement five days. Remote work from Seoul for APAC customers is available at the same rates. All rates in KRW, excluding 10% VAT. Tax invoice (세금계산서) issued for Korean clients; USD invoicing for international clients.
Stack
- Python
- PyTorch
- LangGraph / custom agents
- Claude & OpenAI APIs
- AWS (Bedrock, SageMaker, ECS, Lambda)
- PostgreSQL / pgvector
- FastAPI
- Docker
- Korean NLP & evaluation
Who you get
Usman Asim — AI engineer with four years of production AI work (MSc, Chung-Ang University), plus hands-on network and hardware experience, so on-prem and edge deployments are covered.
Demo
A public demo of a Korean-document RAG agent with a human-escalation queue and dashboard is in progress. Ask for the repository link and a walkthrough video.
Have a Korean customer and no engineer in-country?
Tell us the product, the customer and the timeline. Quote the same business day (KST).