uwcsc / codeybot

Discord CodeyBot used for mock interviewing, suggestions and other features.
18 stars 18 forks source link

[EPIC] Resume support #129

Closed mcpenguin closed 1 year ago

mcpenguin commented 2 years ago

Describe the problem. CSC gets many resumes to roast, often from first years. Currently, Codey does not have any features that helps with the roasting process.

Describe your proposed solution(s). We should integrate some functionality into Codey which somewhat "automates" the resume roasting process. From #270 we already have pdf->image conversion for resumes; the next step should be parsing resumes automatically through some sort of ATS (?) and give some sort of feedback.

EdwinYi-JanYang commented 2 years ago

Scuffed Indeed Job Scraper https://replit.com/@EdwinYiJanYang/WebScraper#main.py

mcpenguin commented 2 years ago

Could we scrape WaterlooWorks? We can keep the jobs secret - it might be nice to just see the proportion of key words that show up for a given "line" of careers (eg Full Stack Dev)

mcpenguin commented 2 years ago

Resume parser code that uses the Lever API: (resume-parser.vercel.app creator) https://github.com/KnlnKS/resume-parser

Guides on how to build a resume parser: https://towardsdatascience.com/how-to-build-a-resume-parsing-tool-ae19c062e377 https://promptapi.com/blog/article/build-your-own-resume-parser-using-python-and-nlp Second one looks really promising, Python might be the best language for this since it is more versatile and has more support for NLP/ML libraries

mcpenguin commented 2 years ago

I'm thinking of registering for Lever's API: https://www.lever.co/partnershipinterest/

There is also https://itsjafer.com/#/parser but I can't seem to make this work :/

mcpenguin commented 2 years ago

In lieu of Lever's resume parser, we could make our own? It seems like Lever made their API even stricter so the request is blocked by CORS once again, so we might want to consider making our own parser instead. The guide I linked last term uses Python, but there should be equivalent TS libraries as well.

@alexz429 @marko-polo-cheno What are your thoughts on this? I don't want to start any R&D unless you guys think this is a feasible idea.

mcpenguin commented 2 years ago

We could also subscribe to the free tier of this API from the same website: https://apilayer.com/marketplace/description/resume_parser-api?utm_source=blog_apilayer&utm_medium=Referral&utm_campaign=blog_banner_apilayer