“I need two data scientists for two months to build a solution” — sometimes this is heard from potential clients. It might work if you know what to do. But usually you don’t know.
In software development, if you had a specification, you could expand the team and do what should be done. In modelling, in fact, you don’t know what’s possible for a long time.
It happened to us that a task was defined as “finding user profiles”, but in fact the client needed a sophisticated recommendation system. How could body leasing help in such a case?
It could work if someone knows exactly what needs to be done and what skills are required. It definitely doesn’t work like “I need two data scientists” to augment a team of software developers and deliver a data product.
It’s risky for the reputation of a company that hires data scientists in this way. Why is that?
Because they won’t see the problem to be solved, but they will be given some tasks to solve a problem as it is seen by a client.
It’s just a reputational risk. Why? In my experience, it’s most convenient to blame a data scientist – not the problem to be solved, not the infrastructure, not the decisions.
These are the reasons why we at QuantUp don’t rent data scientists.
We want to have a part or a whole project under our management and control. And decide how it should be solved and who should do it. We want to take the responsibility, both the risk and the reward.
No, we don’t decide everything without discussing it with the client 🙂
Last but not least, there is an elegant name for body leasing: “team augmentation”. We prefer the former because it describes very well how people are treated in this case.