Data engineer, analyst and scientist used to be three separate briefs. They now arrive as one.

The roles in this discipline are losing their edges. What used to be three separate briefs, a data engineer, a data analyst and a data scientist, now arrives as one. The desk sees data engineers expected to support AI use cases such as RAG pipelines, and analyst responsibilities moving toward engineering and modelling. Some companies have started writing the merge into the job title itself and advertising for a Data Analytics Engineer.

Thailand's own job postings show where the weight actually sits. Of the roles TDRI classifies inside the country's AI job market in Q2 2025, data scientist and data analyst positions numbered 2,618 against 321 for AI and ML engineers, and the data roles grew 23 per cent over the year. The hiring is happening in data, whatever the AI framing on the brief.

The scarce profile is not the one most briefs describe. What we see is that finding someone who knows SQL and traditional ETL is straightforward, and finding someone who has experimented with generative AI is straightforward. Finding someone who can genuinely own the data platform, meaning Python, distributed processing, cloud, data modelling and orchestration, and who also understands how that data is consumed by AI applications, is not. Those two capabilities grew up in different talent pools. Data engineers came from engineering and platform backgrounds; AI people came from data science and machine learning.

The most common failure in a brief is over-specifying the technology and under-defining the problem. A request arrives for a data engineer with Python, Spark, Databricks, AWS, Kafka, machine learning and RAG experience when the underlying need is reliable pipelines and data made available for analytics. The same brief often folds platform engineering, analytics, machine learning and generative AI into one profile without saying where the boundaries lie. This matters more than it sounds: an overly broad brief eliminates strong candidates before the client has spoken to any of them.

Most of the talent comes from large banks, telcos, digital platforms and a small number of Thai technology companies where people have already run modern data platforms at scale. For the stronger data engineering and AI combinations the desk looks overseas, particularly at people with production AI experience on cloud data platforms. The mapping that works is by environment rather than by job title: the same "Data Engineer" title covers very different skill levels depending on what the person actually operated.

When a good candidate says no, the reason is usually not the money. The desk sees hesitation about joining a company with no established Thailand team, because the candidate questions the stability of the operation, the quality of the technical environment and what the career path looks like locally. Compensation can be competitive and the answer is still no if the upside does not cover that uncertainty. Offers have moved accordingly: a higher base, a meaningful annual bonus, and above all hybrid and remote flexibility, which is the term candidates raise most often.

What we tell hiring managers is the same thing four other True Blue desks have arrived at independently: decide what problem the hire will own before writing the brief, and separate the must-haves from the nice-to-haves. And the contrarian read, the one worth acting on: the shortage is of strong data engineers who can adapt to AI, not of people with AI in their job title. Strong engineering fundamentals remain more valuable than generative AI exposure.