Scientific Article Search Engine

Developed an academic search engine for retrieving scientific articles from an arXiv-based document corpus as part of a university project at the University of Béjaïa.

Built the indexing system from scratch using an inverted index, a double polynomial hashing table, and positional posting lists to enable efficient document indexing, keyword lookup, and information retrieval.

Implemented a TF-IDF ranking algorithm to return the most relevant scientific articles for each user query. Also developed a real-time query suggestion module based on Markov Chains to improve the search experience.

Delivered the project as a complete web application: a REST API built with FastAPI, an interactive Streamlit user interface, Docker containerization, GitHub Actions CI/CD pipeline, and deployment on Render.

Optimized indexing, search, and document retrieval performance while maintaining a clean and scalable project architecture.