So, After being laid off from my job at Greenhouse last week, I’ve decided to automate my resume and searching tools. After taking a day or two off to kind of figure out what I really wanted my next steps to be, taking a few days off because of a really nasty stomach bug, and taking a few days to enjoy the winter weather with my puppy, I started to look at my gameplan for figuring out what my next steps are. I will say, that I’ve had some really amazing friends out of all of this:

The Problem

I have a lot of experience. Between my time doing professional services work at CollabNet (yes, even while I was still an Engineer, I still got talked into doing professional services work), my time at Sapient where I was a consultant, and my time where I was doing my own thing, I’ve likely be on-site at over a hundred companies and have had experience doing everything from writing java applications to being the CTO of a small startup. It’s hard for me to always show my best face when I’m writing a general resume for mass review.

My Solution

As I see this solution, I envision it in three distinct pieces:

The Job Parser

Part of this is identifying which positions are great for me…and which ones I probably don’t want to apply for. I plan on writing my own job board parser to digest the job boards of Greenhouse, Lever, Ashby, and whatever else I can find (probably Workday/ADP/Ulti if I can find job board API’s for them). These will look at the open posts and grab the location of the individual jobs. This will then go and scrape all jobs on the job board (and this nightly task will run while I’m asleep, and generally keep the amount of requests down to a sane and reasonable rate).

The Resume Builder

Using a custom system and assistant prompt, I then feed the text of the rendered job description to my LLM. The LLM takes my job history, the job description, and then writes custom descriptions that emphasizes the requirements or skills needed in the job description. Each resume is generated with a unique identifier, to help me refer to it when the recruiter calls or I hop on interviews with teams. I may or may not throw some tracking pixels to indicate when the resume has or has not successfully been parsed (Ron: I told you that dochouse was going to be a problem…ha!)

The CRM

Since I need to know which resume was used when I applied for which specific job, I need to keep track of where it’s been used and which job I’ve applied to. Since most ATS’es send updates, I’ll generate a unique email address for each resume, and automatically take those emails and automatically update the status of each application that I’ve submitted. This abstraction takes the rejection out of the numbers game: if I’ve automatically applied to these positions, and I’m not really investing a lot of time in it, I really won’t care if I’m being rejected from a position.

I don’t know how open source I’m going to make this, but, if there’s enough interest, I may open it up for folks who would be interested in using this tool for their job search. I don’t know how (or if) I could monetize this, so, I’ll probably leave it to ‘friends only’ for a little bit…but, if you’d like to learn more, please reach out to me.