6. LLM Engineer
Evidence label: Specialization or role family, depending on the employer. The title is current, but its scope overlaps ML Engineer, Research Engineer, and AI Engineer.
What they do
Deep specialization in large language model integration and optimization. Where AI engineers are generalists, LLM engineers go deep on the model layer: fine-tuning, RLHF, model selection, prompt optimization at scale, and evaluation infrastructure.
Responsibilities
- Fine-tuning base models for domain-specific tasks
- Designing and running systematic model evaluations (evals)
- Implementing RLHF or similar feedback mechanisms
- Model performance benchmarking and regression testing
- Managing model versions and A/B testing new model releases
- Building tooling for model monitoring and drift detection
Required skills
| Technical | Soft |
|---|---|
| Python (fluent) | Scientific rigor |
| PyTorch or JAX | Statistical thinking |
| Transformers architecture knowledge | Patience with slow feedback loops |
| Evaluation framework design | Documentation of experiments |
| Distributed training basics |
Where it's heading
Strong demand at AI companies (Anthropic, OpenAI, scale-ups) and in large enterprises building proprietary models. Distinct from AI engineer in its proximity to the model itself. Expect this role to bifurcate: pure research at labs vs. applied fine-tuning at enterprises.