LinkedIn Recruiter Outreach Tool

Flow chart of Tool

Inspiration

Wanting to automate the job outreach on LinkedIn; so much time spent looking through pages of recruiters, wanting to find ones that match what I'm looking for.

What it does

Collects mass data on recruiters based on job type without user interaction
Stores recruiter data locally
Data Analytics software to view data sorted by most common location and education - see where your curated recruiters are & where they studied
sort recruiter database by keyword or key phrase; keyword similar match detection
AI generated greeting included for every selected recruiter, curated from user's resume and recruiter's data

How we built it

The recruiter collector was built in Node.JS framework using JS puppeteer. It works by navigating LinkedIn's website and reading html elements to gather data on recruiters. It stores this data in a JSON to be used by the other programs; it is the base of the whole product.

The data analytics software was made in Python with pandas, seaboard and matplotlib.pyplot. It acts as a way to view patterns in your own LinkedIn algorithm; being able to see where your recruiters are located and where they studied.

The sorting functions are in python and javascript; the js script uses direct keyword matching while the python script looks for similar key words/phrases in the data using python's cross encoder and pandas library (ex. if you look for "software developer" it would also find software engineer recruiters).

the generative AI is used in the third program as a part of the suite; its function is to take the selected recruiters and your resume, and generate the best possible greetings for your selected recruiters. The output is used in the front end GUI.

The front end GUI was made in React JS using CSS3 elements - it is a user friendly way to view the selected recruiters and visit their LinkedIn, entering the generated message in and reaching out to them!

Challenges we ran into

Html Elements being dynamic and changing from recruiter to recruiter made it difficult to grab data reliably
file management with several files
working under time constraints

Accomplishments that we're proud of

versatility in frameworks and languages used on this project
data visualization and front end
performance of scraper

What we learned

We learned about the beauty of working with developers that cover the full stack; how we can work together to solve problems
Learned and practiced effective planning, execution and concurrent work, and reflection as a team
Learned more about using GitHub in real time with collaborators

What's next for Recruiter Finder Suite

The next step is bridging the suite into one product; one file. Right now we have the pieces of a great and functional product, the next step is to put them together.