Developers who can easily work with large language models are in high demand in the world of artificial intelligence.

Demand for LLM developers is currently outstripping supply. Even in the world of AI, the technology of large language models is still relatively new and the number of people with the necessary skills is correspondingly large.

money talk

A key factor is financial health, despite recent advances in cost-efficiency. The cost of training an LLM can quickly add up and increase as the model grows. Just look at what it costs to run ChatGPT every day.

polyvalency

In addition, LLM developers must have extensive knowledge of areas such as machine learning, coding, tokens, autoregressive models, transformers, adapters, and reinforcement learning. They must also work in a multi-GPU infrastructure, gather the necessary data to train an LLM, and be comfortable with on-the-fly engineering.

That’s a solid list of qualifications that not many people have. Definitely not enough.

company policy

Potential employers also play a role. Top staff costs money. So if you want to hire a talented LLM developer, you will have to dig deep.

This also includes the necessary support: For example, there must be sufficient possibilities for data collection and for training these AI models. For this, an MLOps team (a combination of machine learning and DevOps) is a requirement rather than a luxury.

So it’s not just about this one developer. Although that is of course the linchpin around which everything revolves.

Given the importance of artificial intelligence, also in the public sector, it can be assumed that there will not be many training opportunities in the near future.