About Us
An independent AI engineering firm. About 50 people. Every one of them an employee.
We're based in Hanoi. We build and prove enterprise AI systems for companies in healthcare, logistics and financial services, mostly in the United States and Europe.
Size
Fifty people is the point, not the limitation.
We can tell you the name of everyone who would touch your data. We can commit to a fixed timebox and a delivery threshold, because there's no approval chain between the person who promises it and the person who does it. We can tell you on a first call that we're the wrong firm for your problem, because nobody here has a quota that punishes saying so.
Firms with three hundred thousand people cannot do any of those things. That's not a criticism of them — it's a structural fact, and it's the reason we compete on the things we compete on.
Leadership
You'll work directly with one of these people, not through them.
Nick Nguyen
CEO / FounderFounder of SyncSoft, with over a decade in IT outsourcing and AI operations. Sets the direction and stays close enough to the work to answer for it.
LinkedIn
Danda Nguyen
Head of AI Data OperationsLeads SyncSoft’s AI data practice — annotation, RLHF and evaluation delivery — and owns the quality process that measures inter-annotator agreement rather than asserting it.
LinkedIn
Jesse Ninh
Engineering LeadLeads engineering delivery across SyncSoft’s agent, LLM-application and integration builds, and is accountable for the eval suites that ship with them.
LinkedIn
Zoe Nguyen
Domain AdvisorSupports projects requiring domain expertise, including healthcare and specialized data interpretation, so the rubric reflects how the work is actually judged.
LinkedInHow We Work
The practical questions, answered up front.
Time zones
Ten to twelve hours of daily overlap with US Eastern and Pacific. Full overlap with Western Europe. Standing daily standup at a time that suits you.
Language
Every client-facing engineer works in English daily and writes the documentation you'll read. The sample evaluation report on our proof page is unedited output — judge it from that.
Tools
Claude Code, Cursor and GitHub Copilot in our own delivery. Your repositories, your CI, your review standards.
What we say no to
Staff augmentation with no defined outcome. Projects where nobody on the client side has time. Fully autonomous decisioning in regulated contexts where we don't think it will pass review. Anything where we can't agree a measurable definition of success.
Two Businesses, One Thread
Why we build agents and produce training data.
Because the loop between them is where the value is. An agent in production generates failure data that nobody else is capturing. That data makes evaluation sharper. Sharper evaluation makes the next system more reliable.
Independence. No AI lab, model provider or cloud vendor holds equity or board representation.
Employment. Everyone on your engagement is a salaried Syncsoft employee — see How We Employ on the trust pages.
Careers. We hire slowly and keep people — ask us our retention number on a call.
Tell us what you're building, or what's broken.
You'll get a reply within one business day, from an engineer. We don't have an SDR team and we're not going to add you to a mailing list.