Baserah AI, Turning Unstructured HR Data into Evaluations

HR teams sit on piles of unstructured text — job descriptions, requirements, evaluation notes — and turning that into measurable competencies and fair assessments is slow manual work. The platform had to extract structure reliably from that text and produce assessments an HR team can actually defend.

• Job descriptions arrive as unstructured text
• Competencies and KPIs extracted consistently, not ad hoc
• Assessments generated, not hand-written per role
• Arabic and English on equal footing

The backend is FastAPI on MongoDB with modular async workflows, using Gemini-powered AI for the extraction and generation steps: competencies and KPIs pulled from job descriptions, MCQ assessments generated from them, and talent evaluated against the extracted structure.

• FastAPI + MongoDB with modular async workflows
• AI extraction of competencies and KPIs
• AI-driven MCQ generation
• Structured, repeatable talent evaluation

An HR pipeline that goes from raw text to structured evaluation.

• Automated competency and KPI extraction from job descriptions
• AI-generated assessments
• Structured talent evaluation workflows
• Localized Arabic/English interface

Baserah turns days of manual HR analysis into an automated pipeline — an example of the applied-AI backend work I build beyond classic CRUD platforms.

• Manual extraction work automated end to end
• Consistent, structured evaluations
• Applied AI in a production backend architecture

Technology Stack