I remember having a lot of fun solving it, in Coffeescript (which was all the rage at the time), also with TSP but using the Pearson correlation coefficient as the similarity metric: https://gist.github.com/stevekrenzel/1505619
Here's a little demo of it in action: http://krenzel.org/files/insta/
In my experience what usually happens is a programmer's "worth" is hugely dependent on the scale of the organization, moreso than nearly every other individual contributor-level role, which means that smaller orgs have an extremely difficult time competing with larger orgs.
For example, suppose at some company there is a business process that generates $1 billion in annual revenue, and a programmer has implemented something that can increase that by, say, .2%. So the programmer's "worth" in that situation is 2 million dollars.
A much smaller company only does $10 million in annual revenue, so the commensurate .2% improvement is only worth 20k.
This is a gross oversimplification, of course, but it really gets to the heart of way the FAANGS can pay such huge salaries and suck up a lot of the best talent. It's not that all the other companies are being stingy, it's just that in many cases they can't pay a programmer anywhere near what a FAANG can because that programmer just can't generate that much business value at a smaller company.
And the number of employees also affects the types of efficiencies you can introduce; the likelihood of finding something that will increase revenue at a higher percentage for a large company is drastically lower than the likelihood of finding something that will increase revenue at a higher percentage for a small company. And that's if it's percentage based in the first place, rather than fixed savings.
Anyway, relatedly, it has more to do with the fact that a lot of the leading tech companies have a high amount of revenue -per employee-. That is, revenue/employee = big number. There's a few different reasons for that, but that's the main thing. Per the example above, it's really more like a company doing $1 billion in revenue has 10k employees, and another doing $100 million has 5k employees. Obviously the former has more to spend on employees, as they're generating more with fewer.
Under that model taken as-is there are some interesting features:
- The company could fire all the engineers and coast if the product is "done". Mostly reasonable, though products are never done.
- More "solve for the equilibrium", engineers are paid for the present value of their contributions rather than some proportion of current revenues. Depends on the company either being well-funded, paying in stock, or being able to take on debt. Again, rings true-ish? Hard to measure so obviously expected present value will differ greatly from the actual value.
There may be a profit shortage among companies claiming they can't find people to hire.
Your company needs to revise what they mean by "average."
Really nasty, lots of headaches, complexity, risks, and/or people think it's magic.