Stanford Graduate School of BusinessQuick Study: Can AI Predict Your Next Job?
Stanford Graduate School of Business
AI Transforms Data Science Work 0:03
Large language models, or LLMs, work by predicting the next word in a sequence, stringing together text that makes sense. The author's team at the Stanford GSB Capital Social Impact Lab wondered whether this same architecture could predict something beyond words, namely a person's future job.
Turning Careers Into Predictable Text 1:31
Using 40 years of worker survey data, the team converted career histories into written resumes, then fed them into a customized open-source LLM. Treating a sequence of jobs like a sequence of words, the model predicted a person's next job more accurately than any other model, a technique also applicable to customer journeys or medical treatment sequences.
Fine Tuning Makes It Precise 2:30
General LLMs know a lot but lack granular accuracy, so the team used fine-tuning to specialize the model, teaching it how the odds of becoming an engineering manager shift as an engineer gains experience. Combining fine-tuned LLMs with rules-based systems and traditional machine learning could soon make businesses, software, and medical predictions more efficient.
AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.
