StartupsAug 19, 2019
GoogletBvm04

Feedback on startup idea

I'm considering building a platform for companies to post real problems they face. Users would compete to win jobs at top companies. Problems would be around robotics, AI, ML and algorithms (but not fizz buzz/leet code questions). The tasks will not be pass/fail, instead, a score would be computed. Example tasks: - given a list of delivery orders, find best route for driver - given a simulator, write Roomba motion planning algo - given sensor data, write localization algo Does this sound fun/interesting? Any feedback appreciated. (I'm looking for a co-founder, must be located in the Bay Area, bonus point for biz/product experience. DM me if if interested or if you know anyone that could be interested)

Add a comment
Apple RoidRage Aug 19, 2019

www.topcoder.com looks like the same thing?

Google tBvm04 OP Aug 19, 2019

It is indeed related. Companies use topcoder to get work done, what I'm working on would be for hiring top talent.

Apple RoidRage Aug 19, 2019

I think some of the leetcode websites also offer to match talent based on skills in solved problems

Uber Hotwheelz Aug 19, 2019

You are conflating two things into one. The need for labor and solution to a problem. Just because you have a real world problem to be solved, does not mean you have open headcount. And vice versa.

Nutanix IIT Aug 19, 2019

The idea is decent, but I share this perspective as well.

Google tBvm04 OP Aug 19, 2019

The problem I am solving is matching talent and companies. The solution is having candidates solve interesting and relevant problems instead of fizzbuzz questions.

Intel HotshotGG Aug 19, 2019

Yeah I like the idea, lets say I solved a problem for my team and I want to see if there is a better way to do it, so I post a similar question on your platform and whoever gives the best answer in terms of efficiency or creativity, they will get a chance at an interview. One of the problems I see is that users might not be motivated enough to solve a problem if it already has many submissions. Another problem is that technical employees will have to scan the answers of the canadiates, because sure as well HR isn't capable of doing it, so this could be very time consuming. But overall I think it is a decent idea and better than all of leetcode questions which doesn't necessary map out to real software engineering, I mean seriously how often do you invert a binary tree in a real job lol

Google tBvm04 OP Aug 19, 2019

Thanks for your feedback, much appreciated. Regarding user motivation, maybe cash rewards would attract users more easily (like Kaggle). Regarding scanning/scoring the results, that can probably be automated.

Chegg vhYe62 Aug 22, 2019

I think quite often the problems are either too big to solve it this way or too specific to the company and their product (like add feature X that integrates nicely with existing architecture).

Red Hat lightvic Aug 27, 2019

Kaggle.com does this for AI/ML