Subscribe to DSC Newsletter

Below is information found today on Kaggle. This flat rate applies regardless of the consultant's age - but only the "best" data scientists are allowed to work on business projects posted by campanies like us (we are going to submit a very interesting project next week). Yet, not sure how much the data scientists are actually paid. In a pure consulting environment, they would get paid 1/3, that is $100/hour - still much higher than rates advertised by recruiting firms.

Yet, if you see Kaggle's demographics, they have tons of users from India, Ukraine or Mexico. If they win projects as frequently as data scientists from other countries, they must be making lots of money: $100/hour goes a long way in Mumbai. My experience is that it is far more efficient to contact Indian experts directly and pay $10/hour, a good rate in India.

Also, are these "best" data scientists really qualified for your project? In other words, is their expensive fee justified, for you? Anyway, my project will consist of proving a new statistical conjecture in the context of non-parametric correlation coefficients for big data. Typically in this type of environment (Kaggle) you find many people with strong theoretical background or great programming skills (frequently overseas), but not many domain experts. I guess it works well for some projects like mine, and not for others.

Kaggle mentions that their "top" data scientists work faster. So maybe they are not expensive after all (likewise, I charge by the project  and not by the hour when I do consulting work, as I usually automate processes and accept only projects that I can do 5x faster than competitors, for half the price).

Here's the extract from Kaggle's website:

Kaggle Connect is a consulting platform that 
connects you to the top 0.5% of the world's data science talent1.

With anywhere from eight to 40 hours, a Kaggle Connect data scientist can help you extract answers and insights from your data2. After you post your project, you will be matched with a data scientist within 60 hours.

To give you the opportunity to fall in love with the service without having to make a financial commitment, we're offering eight complimentary hours for your first small project (no committment or credit card is required). After the eight hours have been consumed, Kaggle Connect costs $300 per hour.

(1) Who are these data scientists?
The Kaggle Connect data scientists are the best 0.5% of the 95,916 data scientists who compete in data mining competitions on Kaggle.

(2) How long is a typical project?
When working with a top 0.5% data scientist, projects take just eight to 40 hours ($3k to $12k). Projects are finished in closer to eight hours for clean data and closer to 40 hours when the data requires cleaning.

(3) How does IP protection and data privacy work?
The data scientist signs a confidentiality agreement before the project begins. All IP is transfered to the company at the conclusion of the project.

Related articles

Views: 2403

Replies to This Discussion


  What has been your experience with Kaggle? I would love to hear the attributes that worked and the ones that didn't. Stillwater is in need of a real-time predictive model in the genomics space but find it very difficult to find biologists that are skilled in modern C++ metaprogramming techniques. So, a question I would have is: how successful is Kaggle connecting domain experts with specific technology stacks, as we are not that interested in a simple answer, we are interested in creating the system that provides the answers.

Love to learn from your experience.


Hi Theodore:

Are you talking about Kaggle competitions? My experience with Kaggle competitions has been disastrous: they just rejected our competition. We ended up doing it ourselves, and it worked very well, with leading mathematicians and professional experts (rather than students or very junior people) participating and winning - the #1 winner ($1,000 award) being Jean-Francois Puget, French IBM distinguished engineer with strong experience in mathematical and combinatorial optimization. But we haven't used Kaggle consultants. Instead, we've hired people in Romania.



On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service