Role vocabulary

ATS resume keywords for AI and Machine Learning

These are the 26 terms our scorer checks a AI and Machine Learning resume against. It is the same list the free check uses, published in full.

The vocabulary

Coverage is 25 of the 100 points. Sixty percent of this list is full marks, so you do not need all of it, and stuffing every term in is penalised separately.

  • machine learning
  • deep learning
  • python
  • pytorch
  • tensorflow
  • scikit-learn
  • nlp
  • computer vision
  • llm
  • transformer
  • fine-tuning
  • rag
  • embeddings
  • feature engineering
  • model training
  • model evaluation
  • mlops
  • inference
  • hugging face
  • pandas
  • numpy
  • data pipeline
  • experimentation
  • accuracy
  • vector database
  • deployment

The eight questions for AI and Machine Learning

The guided builder asks these, in this order, and turns each answer into a line on the page. Nothing is written that you did not type.

  1. 1.How many models have you shipped to production, and what do they do?
  2. 2.What accuracy, precision, recall or business metric did your best model move, and by how much?
  3. 3.How large are the datasets you train on, in rows, tokens or gigabytes?
  4. 4.Which frameworks, libraries and platforms do you use daily?
  5. 5.What does your model serve in production, in requests per day or users?
  6. 6.How much time, cost or manual effort has your model or pipeline saved?
  7. 7.How many engineers, analysts or stakeholders do you work with or lead?
  8. 8.What is the hardest data or modelling problem you have solved, and how?
Answer them and see your score before paying

Words are the easy half

A resume can carry every term above and still be rejected, because coverage says what you know and nothing about what you delivered. The heaviest category on the rubric is evidence: how many of your bullets carry an actual figure.

Seven in ten is full marks. Most resumes we see manage two.

Get your number free

Other functions