The AI engineer's job is splitting in two, and most briefs are still written for the old one.

The engineer's job is splitting in two. What clients asked for this year no longer resolves to one role. Forward deployed engineers sit in front of the customer and need consulting instinct alongside deep technical skill. AI product engineers are a hybrid of software engineer and product manager. As AI tools get better at coding, an engineer is no longer required to spend all day producing lines of code, so the job moves toward the product and the customer. Alongside both, a third strand is pure AI system design, where retrieval augmented generation and strong Python are the skills in demand right now.

Which is why most briefs we receive are asking for the wrong thing. Forward deployed engineering is a very new role in Thailand. It is not going to be on anyone's CV yet, so expecting masses of commercial experience in it rules out every candidate who could actually do it. The assessment has to be about crossover skills. The supply side backs this up: Thailand's large-scale national AI bootcamps offer around 10,000 places a year and produce about 2,000 completions, so the trained pool is small and the demand is not.

The hardest thing to find is not the engineering. For forward deployed roles it is English strong enough for a regional client-facing job, paired with the consulting temperament to use it. Engineers have traditionally skewed introverted, so the ones who are comfortable in front of a client are a narrow slice of an already narrow pool. The other constraint is adaptation. A large number of otherwise strong candidates are working in legacy tech or in an old-fashioned way, and that is not what anyone is buying. Clients feel this from the other side: 51 per cent of Thai organisations name a shortage of qualified technology professionals as their primary obstacle to AI adoption, ahead of cost and regulation.

When a good candidate says no, it is often not about money. It is about the work not being at the cutting edge. That said, money is real: AI firms are paying market-leading packages, and organisations that have not prepared for that lose people late. On the searches our desk is running, strong engineers with proven production AI experience are commanding THB 150,000 to 250,000 a month, with bonus and equity on top at the upper end. That band is what we see on live offers, not a measured market figure.

Where the people actually are. Startups and innovators sit closest to the frontier and corporates tend to lag, so that is where a lot of our strongest candidates come from. We are also increasingly looking internationally, mapping Thai nationals working abroad with cutting-edge skills who want to come home, because Thailand still sits behind markets like the US on genuine frontier knowledge. The public data agrees: Thai workplace AI use grew 36.4 per cent in the year to Q1 2026, second fastest in the world, but the absolute rate of 12.4 per cent is still below the global average of 17.8 per cent. Fast growth, low base.

What we tell hiring managers. If you cannot pay at the top of the market, be prepared to nurture rather than buy. Look past the hard skills to fundamentals, a willingness to learn, and genuine engagement with AI tools. Someone mouldable with fresh ideas will outperform twenty years of experience that refuses to change how it works. The underlying shift is the one to plan for: the software engineer is not dying, it is evolving, and what the title means in five years will not be what it means today. That shift is already visible and it will accelerate.