Skip to content

Instantly share code, notes, and snippets.

@danking
Created February 5, 2020 20:07
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save danking/3d6fac3ffd9419b50343712365c9bca0 to your computer and use it in GitHub Desktop.
Save danking/3d6fac3ffd9419b50343712365c9bca0 to your computer and use it in GitHub Desktop.

Co-op Resume Filtering

Background

In recent semesters, Hail has received >300 applications for our one open co-op position. We engage in an uncomfortable, severe filtering process to cope with this overwhelming number of applications.

For the Spring 2020 position, Hail received 316 applications. For the upcoming Fall 2020 position, Hail has thus far received 162 applications.

When the team filters, we expend a tremendous amount of person-hours. A one hour filtering with ten people is more than a person-day of productivity lost.

The Process

I intend to spend no more than one hour on filtering.

We are likely to receive around 200 applications. We can only hire one person. We will interview five or fewer people. The mean time we'll spend on a resume is 200/60 ~= 3.5 minutes. In years past, there has been a subjective-quality-cliff that separates as many as 90% of the resumes from the 10% of high-quality resumes. Let N be the number of hail team members present.

Bring laptops.

Step 1

15 minutes

Divide the applications evenly into N groups. Each person chooses roughly the top 25% from their pile. If your pile has 20 people, you should aim to advance five or fewer resumes. Advance no more than seven (35%).

If there is a group of three clearly superior resumes and distinguishing the two next-best resumes is substantially more difficult, then simply advance three.

Step 2

15 minutes

You should have no more than 35% of your total remaining. Again, divide the applications evenly into N groups. Set the target percentage such that between fifteen and twenty-five resumes will remain. If there are ten hail team members and 70 resumes remaining, each team member will get seven resumes and should advance between zero and three resumes. Aim to advance two resumes.

Step 3

The leader emails the remaining resumes to the team. Each team member gives a ranked choice of their top five resumes. For example, if Jackie's top choice is Claire and her fifth choice is Zan, Jackie should email the leader an email containing exactly this text and nothing else:

Claire Ben Hilary Dan Zan

The leader then conducts a Single Transferable Vote election using the simplest method given on Wikipedia: https://en.wikipedia.org/wiki/Single_transferable_vote#Simplest_method:_elimination_transfers_only

The suggested amount of time for step 3 is fifteen minutes and no more than thirty minutes.

Resume Evaluation

We are looking for candidates that:

  • can quickly and correctly solve well-defined programming problems

  • will collaborate positively with us

  • bring a new skill set or perspective that we don’t have

The latter is hard to find in a co-op, but keep your eyes open!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment