Hours of CV screening, cut to minutes
The HR team was spending four to six hours per batch reviewing, translating and re-typing CV data by hand. ISM built an LLM pipeline in Python and LangChain that ingests the documents, translates them, extracts structured fields and validates every record against the HR schema before it lands.
It shipped to production during the engagement and the HR team runs it today.
Discuss something similar| Measure | Before | After |
|---|---|---|
| Time per batch | 4 to 6 hours | Under 5 minutes |
| Handling | Manual re-typing | Extracted and validated |
| Languages | Translated by hand | Detected and translated |
| Status | Spreadsheet | In production with HR |
Figures are the project’s own measurements, taken before and after the pipeline went live.
