My name is Nick, and I'm a hobbyist programmer! Several years ago after applying to hundreds of jobs began to recognize that companies had been using tools to help make the hiring process more efficient from their end and realized candidates lacked really any tools similar so decided to make one! I figured that, like me, other job seekers were stuck manually applying to online jobs by repetitively filling out the same information over and over again with really no good tools to do it for them.
So, I built AA (aka AutoApply)... and it's almost finished!
AA is a 1000% free, fully autonomous, open-source agent that discovers, vets, and applies to jobs for you. Here is how the program is broken down and works
Discovery: It reads and extracts information from search engines, job boards, and company career pages to find available jobs.
Vetting: It utilizes local NLP (SpaCy) to ensure a "Two-Way Fit" meaning is this a job the candidate finds fit and does the candidate align with what the company is looking for. It checks if attributes like salary, location/commuting distance, and required skills match your profile information!
Apply: It can upload resumes, cover letters, CV's, handles multi-page forms, logins, and even uses a local offline AI (GPT4All) to write personalized answers to open-ended questions such as "Tell us about a challenge you over came at specifically 6 years old and why you didn't graduate from Harvard".
I designed this with a few core philosophies in mind that I'm really proud of
-Worst-Case First: It designed to be run on a 2GB RAM library computer with zero admin rights.
-Flash Drive Capability: I don't own nor can afford a laptop so can download AA onto a USB stick(flash drive) and run it completely portably.
-Undetectable: It has a built-in evasion framework. It spoofs lots of browser information and human behavior (Bezier-curved mouse movements & parabolic typing delays).
-Accommodate Everyone: I hate when things don't work so throughout the entire build I always wanted to accommodate to every and all users NOT them accommodate to or for me!
-Everything Customizable: I mentioned it can work autonomously but can be run other ways such as with user supervision "Human-in-the-Loop (HITL)" which means you can set checkpoints for where you wish AA to stop to input specifically personal information if you'd like.
-The Research Platform: Nearly all hiring/job market data is owned and released by employers. AA has an opt-in, strictly anonymized (Zero-PII) research module to perhaps offer a new perspective on that data. It recognizes and reports things like ghosting rates, biases, various forms of discrimination, unethical business practices, and plenty more. My goal was and is to build a candidate-side dataset on the hiring market that is academically publishable.
I'm almost done with the project so if you're job hunting and think you could/would use this please leave a note or comment! If you are a developer/engineer, I'd love your feedback on anything from the architecture (strict Hexagonal Ports & Adapters design) to the safety and security production level code for a project like this and even just professional insight would be amazing! If you are a Mathematician, my Math Subsystem utilizes spatial geometry to break down webpages for the autonomous feature and I'd like all my Math to be deterministic, pure mathematical algorithms for the research and "open-ended questions" code and finally my Research Signal Detectors use things like Wilson's score interval, Fisher's exact test, Kendall's Tau, etc.! If you are a Researcher or in Academics, I truly have no idea what I'm doing nor if anything is correct so any review or help or insight would be phenomenal!
Liebmann5•1h ago
My name is Nick, and I'm a hobbyist programmer! Several years ago after applying to hundreds of jobs began to recognize that companies had been using tools to help make the hiring process more efficient from their end and realized candidates lacked really any tools similar so decided to make one! I figured that, like me, other job seekers were stuck manually applying to online jobs by repetitively filling out the same information over and over again with really no good tools to do it for them.
So, I built AA (aka AutoApply)... and it's almost finished!
AA is a 1000% free, fully autonomous, open-source agent that discovers, vets, and applies to jobs for you. Here is how the program is broken down and works
Discovery: It reads and extracts information from search engines, job boards, and company career pages to find available jobs. Vetting: It utilizes local NLP (SpaCy) to ensure a "Two-Way Fit" meaning is this a job the candidate finds fit and does the candidate align with what the company is looking for. It checks if attributes like salary, location/commuting distance, and required skills match your profile information! Apply: It can upload resumes, cover letters, CV's, handles multi-page forms, logins, and even uses a local offline AI (GPT4All) to write personalized answers to open-ended questions such as "Tell us about a challenge you over came at specifically 6 years old and why you didn't graduate from Harvard".
I designed this with a few core philosophies in mind that I'm really proud of
-Worst-Case First: It designed to be run on a 2GB RAM library computer with zero admin rights. -Flash Drive Capability: I don't own nor can afford a laptop so can download AA onto a USB stick(flash drive) and run it completely portably. -Undetectable: It has a built-in evasion framework. It spoofs lots of browser information and human behavior (Bezier-curved mouse movements & parabolic typing delays). -Accommodate Everyone: I hate when things don't work so throughout the entire build I always wanted to accommodate to every and all users NOT them accommodate to or for me! -Everything Customizable: I mentioned it can work autonomously but can be run other ways such as with user supervision "Human-in-the-Loop (HITL)" which means you can set checkpoints for where you wish AA to stop to input specifically personal information if you'd like. -The Research Platform: Nearly all hiring/job market data is owned and released by employers. AA has an opt-in, strictly anonymized (Zero-PII) research module to perhaps offer a new perspective on that data. It recognizes and reports things like ghosting rates, biases, various forms of discrimination, unethical business practices, and plenty more. My goal was and is to build a candidate-side dataset on the hiring market that is academically publishable.
I'm almost done with the project so if you're job hunting and think you could/would use this please leave a note or comment! If you are a developer/engineer, I'd love your feedback on anything from the architecture (strict Hexagonal Ports & Adapters design) to the safety and security production level code for a project like this and even just professional insight would be amazing! If you are a Mathematician, my Math Subsystem utilizes spatial geometry to break down webpages for the autonomous feature and I'd like all my Math to be deterministic, pure mathematical algorithms for the research and "open-ended questions" code and finally my Research Signal Detectors use things like Wilson's score interval, Fisher's exact test, Kendall's Tau, etc.! If you are a Researcher or in Academics, I truly have no idea what I'm doing nor if anything is correct so any review or help or insight would be phenomenal!