git push push push sahur

Users should use our UI to input LinkedIn URLs. Once entered, we’ll do a full account scrape, parse the html returned, and run the html through an LLM. We cache the output of the LLM for query speed in the future.

Quality of life features include robust filtering by full-text search, location, companies, skills, and profile completeness. We also have an analytics page with various graphs and charts, providing the user even more insights. We also implemented a background job which refreshes all the users cached data, so that users may get the latest alumni information.

We used Selenium and Beautifulsoup for scraping. We used a Google LLM for structuring data. All of this backend is powered by Python. The frontend uses NextJs and ShadCn for a robust user interface.