Tasks for NLP Engineers
Tasks are the standard problems NLP systems are built to solve, each with its own output format and metrics. They include text classification, named entity recognition, question answering, summarisation, machine translation and semantic similarity. Choosing the task framing decides the model head, the data labels and the evaluation method.
Stage: Adaptation and Alignment · Part of the NLP Engineer roadmap · Last reviewed 19 November 2026
What to learn in Tasks
- Frame a business problem as classification, tagging or generation
- Label tokens for named entity recognition with BIO tags
- Build an extractive question answering pipeline over a passage
- Compare abstractive and extractive summarisation approaches
- Pick a suitable metric for each task type
Tasks interview questions
What are BIO tags in named entity recognition?
BIO tags mark each token as the Beginning of an entity, Inside an entity, or Outside any entity, usually with a type such as B-PER or I-LOC. This lets a token classifier recover multi-word entities and their boundaries.
How does extractive question answering differ from generative answering?
Extractive models read a passage and predict the start and end positions of an answer span, so output is always copied text. Generative models write the answer token by token, which is more flexible but can invent unsupported details.
Where Tasks fits in the NLP Engineer roadmap
Tasks belongs to the Adaptation and Alignment stage (advanced level), which takes about 41 hours. Make models useful.
Other units in this stage
Back to the full NLP Engineer roadmap