Offshore Development Center in Vietnam

An offshore development center (ODC) is a team of engineers in another country that works only on your product, on your roadmap, for as long as you need it. The vendor employs the team and runs the people side: hiring, payroll, office, equipment and retention. You set priorities and make product decisions.

MOHA Software runs ODC teams from our office in Hanoi (Level 8, AZ Office, 72 Dich Vong, Cau Giay). Our Japanese entity, MOHA Japan LLC, has operated from Yokohama since January 2025. Much of our delivery work has been for Japanese clients, and that shaped how we document, test and report.

This page covers what an ODC is good for, when to pick something else, the team shapes we have actually shipped with, how setup and governance run, what to put in the contract, and what drives cost.

ODC, fixed-price project or staff augmentation?

Three models get sold under many names. The difference comes down to who owns what.

In a fixed-price project, the vendor owns scope and delivery against a signed spec. In an ODC, you own the backlog and the vendor owns the team. In staff augmentation, individual engineers join your team and your managers run them day to day.

An ODC is the right call when the work is ongoing and you want the same people to build up knowledge of your codebase over months and years. If you have a one-off build with a clear spec, a custom software development project with fixed scope is simpler. If you need one specialist for a quarter, staff augmentation is cheaper to set up and cheaper to end. We compare the last two in more depth in staff augmentation vs full outsourcing.

Who an ODC fits, and who it doesn’t

It tends to work when:

Your product will need engineering capacity for a year or more.
Someone on your side (CTO, VP Engineering, head of product or a senior tech lead) can set priorities and review work every week.
Hiring locally is slow or expensive, and you need a team rather than one or two people.
You want continuity. Rotating contractors lose context every time someone leaves.

It usually fails when:

Nobody on your side owns the backlog. An ODC without a product owner drifts, wherever it is located.
You need a fixed deliverable by a fixed date on a fixed budget. Buy a project instead.
Your work legally has to stay in your country. Check data residency and export rules before you speak to any offshore vendor.

Team shapes from projects we’ve shipped

Most ODCs start small and grow once the working rhythm is stable. Rather than a generic org chart, here are team sizes from our published case studies:

Project	Team	Duration	Stack
COMY language learning LMS	6: 1 PM, 2 backend, 1 frontend, 1 QA, 1 UI/UX (27.49 man-months)	About 5 to 6 months	Laravel, MySQL, jQuery, Bootstrap
Hachi:um online teaching platform	6	5 months	Laravel, MySQL, AWS
Laos Edu learning, testing and e-library	10	6 months	Angular, Spring Boot, MySQL, React Native
StarHub customer support chatbot	16	Not published	C#, Python, Docker, GCP

The COMY shape is a common starting point for a product team: one person managing delivery, a small engineering core, dedicated QA and a designer. Teams grow from there when the backlog justifies it, usually by adding engineers before adding management.

Beyond these projects, the stacks we list on our service pages include React, Angular, Vue, Node.js and NestJS, Java, .NET, PHP and Laravel, Flutter, Kotlin, Swift, and AWS, Azure and GCP, with Docker, Kubernetes, Terraform and GitLab CI on the delivery side. Ask which of these we can staff now for your start date, since that changes month to month.

How setup runs

Our engagements follow five stages.

Discovery. We learn about your product, users, stack and the problem the team should solve. It starts with a conversation about what you need.
Requirement definition. We agree on what the team will work on first and what “done” means.
Team setup. We pick people with the right skills, experience and seniority for your backlog.
Project start. Access to your repositories and tools, a walkthrough of your architecture, first small tickets.
Performance tracking. We monitor performance, quality and milestones to keep the work on track.

Plan for the first month to be slower than the ones after it. New engineers need time with an unfamiliar codebase and domain, however senior they are. Judge a new ODC on months two and three.

How governance works

You own priorities. The vendor owns making sure the team delivers against them, and telling you early when something slips. A well-run ODC looks like this:

The team works in your tools where you prefer: your Git host, your Jira or Linear board, your Slack or Teams, your CI pipeline.
Your ceremonies, your cadence. Sprint planning, stand-ups, reviews and retros run on your calendar.
Code review on every pull request, test cases owned by QA, and a regression pass before each release.
A written progress update that covers what shipped, what’s blocked, risks and team changes, so nothing important depends on a single meeting.

One example of how we approach QA is our internal platform, Test-case AI, pairs a library of more than a hundred test-case templates written under ISTQB guidelines with a fine-tuned LLM that drafts test cases from specs and mock-ups. It came out of a practical problem: writing cases from scratch every sprint drained time, and coverage of edge cases was inconsistent. Certifications held across our team include AWS Certified Developer, Solutions Architect and SysOps Administrator (Associate), PMP, Certified ScrumMaster and ISTQB.

What to put in the contract

Every CTO asks the same questions before signing. Whoever you hire, make sure the paperwork answers them in writing:

IP assignment. Code, designs and other work product created for you are assigned to you, stated explicitly in the master services agreement.
Confidentiality. An NDA between the companies, plus confidentiality obligations on every individual engineer through their employment terms.
Access. Engineers use accounts you issue and can revoke. Use SSO and least-privilege access, and keep production data out of development environments.
Replacement. What happens when someone leaves or isn’t working out: how fast a replacement arrives, how handover works, and who pays for the overlap.
Scaling and exit. Notice periods for adding or removing people, and a handover plan if you end the engagement or move the team in-house.
Security questionnaire. Send yours early. A vendor that takes weeks to fill it in will be slow on other things too.

What drives the cost of an ODC

We don’t publish rates, because the cost of a team depends on its exact mix of roles and seniority, and a generic number misleads more than it helps. What we can do is explain what the number is made of, so you can compare offers properly.

Compare fully loaded cost, not salaries. In the US, benefits alone accounted for 30% of what private employers spent on compensation in June 2026, according to the Bureau of Labor Statistics Employer Costs for Employee Compensation release (market data, all private-industry workers). A base salary covers only the remaining 70%, and that’s before recruiting fees, office space, equipment, management time and the cost of replacing people who leave.

On the ODC side, ask what the monthly fee includes. The usual drivers are:

Seniority mix. A team heavy with seniors costs more per head and usually needs less of your time.
Roles beyond engineering. PM, QA and design time are real costs, whether billed separately or bundled.
Ramp-up and overlap. Onboarding time and any overlap during replacements.
Tools and infrastructure. Licenses, cloud environments and devices, and who pays for each.
Your own time. A product owner on your side is a cost too, and an ODC won’t work without one.

Tell us the team you have in mind and we’ll discuss it with you.

Why run your ODC in Vietnam

Vietnam ranked 7th of 78 countries in Kearney’s 2023 Global Services Location Index, which scores locations on financial attractiveness, people skills and business environment for offshore services. Hanoi and Ho Chi Minh City have large pools of web, mobile, Java, .NET and PHP engineers. Senior AI specialists and niche platform experts are harder to find, so expect longer lead times for those roles anywhere in the country.

Hanoi runs on GMT+7 with no daylight saving. For teams in Singapore and Australia the working days mostly overlap. London gets the Hanoi afternoon. US teams get a thin slice at the edges of the day, which works with one fixed daily call and good written handoffs. Our guide to Vietnam offshore software development goes further on the country itself, and we’ve written about how AI changes ODC team design.

FAQ

How is an ODC different from a dedicated team?

Vendors use the terms loosely. An ODC usually means a long-term team that works only on your product and grows over time, with the vendor handling HR, office and retention. “Dedicated team” often describes a smaller or shorter setup. The day-to-day working model is the same.

How big should a first ODC team be?

Big enough to ship without you filling gaps. Our COMY LMS team (1 PM, 2 backend, 1 frontend, 1 QA, 1 designer) is a typical first shape. Starting with four to six people and growing is easier than starting with fifteen and shrinking.

Does an ODC in Vietnam work for a US company?

Yes, if you design for asynchronous work. Keep one short daily overlap call, write decisions down, and give the team enough context that they aren’t blocked waiting for your morning. If your team needs constant live pairing across the whole day, the time difference will hurt.

What should I ask for before signing?

Written IP assignment, individual confidentiality terms, a replacement process with timelines, scaling and exit terms, and a completed security questionnaire. Also ask to see profiles of the actual people who would join your team.