Pitch: Finding the best development talent is key to any startups success. It can be difficult to properly interview developers even when you have a strong technical background. But Github never lies.
Backstory: This is one of my favorite ideas that we never pursued. We actually had some great early validation in terms of potential users and interest from investors. The major holdup that kept us from making this a reality was the technical difficulty. While I'm aware of how difficult it will be, the potential upside is incredible. Being able to accurately compare and evaluate a developer's skillset based on the actual work they've done is the best way to identify technical talent.
Automated Github Reviews
At the highest level, the idea is simple. A developer allows Automated Github Reviews (AGR) access to their Github account. AGR then crunches the data and evaluates the developers talent and proficiency with different languages and overall as a developer.
There's such a treasure trove of data you can easily identify from Github without much effort. How many different repositories have they worked on? How large are these projects? How many projects have they worked on with various programming languages? How many commits have they made in the last month? year? These are the easy ones.
Then you can get into a lot more difficult data analysis. Code climate does a very good job of this on a single project basis. There's a lot of overlap with what they do to ensure code quality on a project level with what could be done to evaluate talent.
Tech companies are currently paying tens of thousands of dollars PER developer that they hire to recruiting agencies, internal recruiters, and interviews and code reviews. There's clearly a massive revenue opportunity if you can figure out a way to automate the assessment of a Github user profile.