Hiring FDEs in Asia: language and years with customers are hard requirements

You open a Forward Deployed Engineer job ad in Tokyo, read the job description and feel you fit. Then, under requirements, you hit one line: fluency in both Japanese and English, spoken and written. You click through to another company’s ad, and the page says the role has been filled.

One point needs stating at the outset: most of the ads mentioned below have been taken down, have stopped accepting applications or have been filled. Read them as a snapshot of the companies that have hired FDEs in Asia-Pacific, not as a list of roles you can apply for today.

The value of that snapshot lies in the requirements. The skill to practise is reading a job ad the way you would read a customer’s requirements document: separate the hard requirements from the description, then check yourself against them honestly.

Who has been hiring, and where?

Tokyo has the biggest names. OpenAI describes its Tokyo FDE as someone who leads complex, end-to-end deployments that take frontier models into production. The role is hybrid, three days a week in the office, comes with relocation support and involves travel mainly within Japan.

OpenAI also once posted a more senior Tokyo role, “Forward Deployed Software Engineer”, requiring 7+ years of full-stack experience and fluency in both languages; that ad was later taken down.

Anthropic has also hired in Tokyo for a “Forward Deployed Engineer, Applied AI”, who works directly alongside its most strategic customers to build production applications on Claude. The ad was taken down in April 2026.

Singapore is more varied. OpenAI posted an FDE role there requiring 4+ years of experience and a willingness to travel up to 50%, working on site with customers; its careers page now says it is no longer accepting applications.

A foundation-model startup once hired an FDE based in Singapore, working remotely or from the office and covering customers and partners across Asia-Pacific; the ad was taken down on 24 February 2026.

Demand in Singapore does not come only from LLM companies either. FieldAI, a robotics company, has hired FDEs to lead end-to-end deployments of autonomous systems in customers’ own environments.

Singtel Group posted an FDE role on 6 May 2026, and its careers page now shows it as filled. Anthropic has also announced an office in Singapore, its fifth in Asia-Pacific, with open roles that include applied AI.

Seoul has Notion. The company once hired an FDE (GTM - Korea) on a hybrid basis in Seoul to lead technical projects with customers, requiring native-level Korean and professional working English, spoken included. The ad was taken down on 3 July 2026.

Where are the hard requirements?

Put these ads into one table and you will see that each market screens out applicants at a different point.

Market	Example companies	Hard requirements	Working pattern
Tokyo	OpenAI, Anthropic	Fluent Japanese and English; 5+ years (OpenAI) or 4+ years (Anthropic) of technical work with customers	OpenAI: hybrid, 3 days in the office, travel mainly within Japan. Anthropic: 25-50% travel to customers
Seoul	Notion	Native-level Korean, professional working English	Hybrid, leading technical projects with customers
Singapore	OpenAI, FieldAI, Singtel, foundation-model startup	OpenAI: 4+ years of experience, willing to work on site with customers	OpenAI: up to 50% travel. Foundation-model startup: covers the whole region

In Tokyo and Seoul, language is the first barrier. The reason is easy to see: an FDE sits with the customer, reads their internal documents and meets their operations teams. If you cannot work in the customer’s language, the role is hard to do, so employers list it as a requirement, not a nice-to-have.

Singapore is different. The barrier there is travel: an FDE based in Singapore may have to look after customers in several countries, and when an ad says up to 50% of time on customer sites, that is half a year away from your own desk.

An example: turning a job ad into a matrix

Picture a backend engineer with 5 years of experience, good English and intermediate Japanese. Of those 5 years, about 2 were spent on integration work directly with enterprise customers; the rest went into building internal services.

Set OpenAI’s Tokyo ad against that profile. It asks for 5+ years of technical work with customers, not 5 years as an engineer.

So this engineer has 5 years in the profession but only about 2 of the kind the ad counts. Add the bilingual requirement and the profile has two gaps, both of them hard requirements.

Anthropic’s Tokyo ad lowers the bar to 4+ years of technical work with customers, but still requires fluent Japanese and English, plus 25-50% of time travelling to customers. The gap is smaller but still there.

For OpenAI’s Singapore ad, the standout conditions are 4+ years of experience and a willingness to work on site with customers up to 50% of the time. Here the key question for this engineer is no longer language but whether they will accept that pace of travel.

From this matrix the path is fairly clear. In the short term, Singapore is the more realistic way in for this profile. Tokyo needs longer preparation: bringing Japanese up to working level, and actively taking on more customer-facing work at the current employer to build up the years.

How to do this for your own city

Start with a list of companies by city, not a list of job ads. Many ads in this article are closed or filled, but a company that has hired an FDE in a city usually has customers there. Watching a company’s own careers page is more reliable than waiting for ads to show up on aggregators.

For each ad, write out four columns: language, years, type of experience (general engineering, full-stack, or work with customers) and amount of travel. Mark which columns are hard requirements and which are only preferences. Then write your own figures alongside, without rounding up.

Finally, turn the matrix into a CV. If the ad counts years with customers, put that number first and tie it to concrete work: which system, for which customer, and how it ran in production.

The phrase “end-to-end deployment” appears in both OpenAI’s and FieldAI’s ads, so each bullet point should tell a full cycle, from receiving the requirement to the system running for real. If you are aiming at Tokyo or Seoul, state your language level against a standard employers can verify.

Which mistakes get applications rejected?

The most common mistake is counting the wrong kind of experience. Three years building internal APIs are not three years working with customers, and interviewers will probe exactly that point.

Close behind is assuming that “decent Japanese” is enough when the ad clearly asks for fluency in both speaking and writing. Applying before you meet the bar rarely pays off, so the effort is better spent on learning the language.

Many people also look only at LLM companies. FieldAI and Singtel show that robotics and telecoms in Singapore also need people to take systems into customers’ real environments, so if you have worked on infrastructure or embedded systems, those are places worth trying.

The last mistake is treating a closed ad as a closed door. A filled role may mean a team is growing, or simply a replacement hire, but either way it shows the company needs FDEs in that city.

What matters most is knowing who that team will need for its next hire, and having your application ready against exactly those requirements.