Our Matching Process

Skill translators only scratch the surface

Most skill translation software relies on the common skills associated with a particular military job title. That means the opportunities look exactly the same for anyone with your military job.

Military skills chart

They lump you into groups based on occupational specialties, which prevents them from having to do the hard work of understanding you.

You deserve better.

How we find the perfect match

Shift matches great people to great organizations. We’re building matching technology to assess these attributes:

  • Career preferences
  • Behaviors
  • Practical skills
  • Personality traits
  • Workplace culture fit

We believe your experience and interests can align to a job that pays you the most and makes you the happiest. Here’s a sample of output of candidate preferences:

Our partner companies want to hire a great fit and we work closely with them to understand the culture of their organization and hiring needs to continuously improve our matching process.

Getting technical: creating vectors and computing distance

Once a candidate completes our assessments, we compare their profile to employer profiles we generate from surveys complete by all of our partner companies.

We create a vector for each candidate — one value for each dimension we care about. For example, if a Shift candidate indicates that she is particularly interested in sales roles, then her "sales" dimension would receive a large value. If she prefers small organizations, then her "organization size" dimension would receive a low value.

Here's what a sample candidate role vector might look like:

 

Individual vector chart

 

We create a similar vector for each role at each of our partner companies. The dimensions are comparable between roles and prospective candidates.

Here's what a sample company's role vector might look like:

 

Role vector chart

 

Each vector describing a person or a role has dozens of dimensions. It's tough to visualize more than three to four variables, so we've used a technique called principal component analysis to smush about 50 dimensions into three: x, y, and z, which we can visualize on a 3D graph.

When we match, we use a distance function to generate a ranked list of candidates for each role. You can think of this step as drawing a line between blue and green dots.

  • Each blue dot represents a role at one of our partner companies.
  • Each green dot represents a fellow (this is some of our actual fellows data, though it's anonymized).
  • The shortest line from a person to a company indicates the strongest match.

We're constantly iterating, so the specifics of our matching technology will certainly evolve. If you’re interested in learning more please reach out to data@shift.org.

Ready for a civilian career that fits you like a glove?

Don't find a job. Find the job.

Apply Now