← Back to projects
In ProgressGitHub
CareerPath-AI
FastAPI backend that ingests a student profile (courses, GPA trend, projects, certifications) and returns top role recommendations with model scores using a serialized ML model.
PythonFastAPIMLLightGBM
Overview
CareerPath-AI is an in-progress recommendation service built around a FastAPI API. It accepts a StudentProfile schema (courses, GPA trend, projects, certifications, extracurriculars), converts it into feature vectors (counts plus GPA mean and slope), and scores roles using a joblib-loaded model stored at data/models/careerpath_lgb.pkl. The API exposes a health check and a /api/recommend-role endpoint that returns the top three roles with probability scores.
Highlights
- Pydantic schemas define StudentProfile inputs and RoleRecommendation outputs.
- Feature engineering computes counts, GPA mean, and GPA trend slope.
- Model loader uses joblib with label-aware probability mapping.
- FastAPI router exposes /api/health and /api/recommend-role endpoints.
- ML toolchain includes LightGBM, CatBoost, transformers, and SHAP for experimentation.