I was into HackerRank for a while (in a previous life I did a lot of TopCoder/CodeJam/ICPC/USACO/etc) but it just wasn't the same. Programming competitions have a long history and high standards, and there were just too many repeated failures on both the side of execution and problem selection. An objective function that immediately allows a trivial solution for record points. Lack of compiling with optimizations. Changing what libraries are allowed partway through (more accurately, some server instances allowed the library, but they all went away by the end). Too basic algorithms. That said, I did have a lot of fun, especially having free reign to use ML and NLP and math libraries. But even that just made me realize that ML competitions tend by nature to be less about the "Aha" moment then they are about getting all your ducks in a row and doing the appropriate due diligence, and I might as well spend that time consulting for hedge funds.
Comments
I was into HackerRank for a while (in a previous life I did a lot of TopCoder/CodeJam/ICPC/USACO/etc) but it just wasn't the same. Programming competitions have a long history and high standards, and there were just too many repeated failures on both the side of execution and problem selection. An objective function that immediately allows a trivial solution for record points. Lack of compiling with optimizations. Changing what libraries are allowed partway through (more accurately, some server instances allowed the library, but they all went away by the end). Too basic algorithms. That said, I did have a lot of fun, especially having free reign to use ML and NLP and math libraries. But even that just made me realize that ML competitions tend by nature to be less about the "Aha" moment then they are about getting all your ducks in a row and doing the appropriate due diligence, and I might as well spend that time consulting for hedge funds.