Analytics is slowly vanishing as a term, not that it is getting old. We are getting into advanced mode and hence the birth of terms like Data Scientist, Artificial Intelligence. Here we are busting some myths about Hiring Data Scientists.
- All Data Scientists come with experience & flashy Degrees/PhDs – well, not necessarily, Data scientist is a recent development and folks with experience don’t really matter. I have hired DSs( you need to keep up with the acronyms, saves typing time you see). Yes, I have hired DSs without PhDs and with 2-3 years of experience & these folks are good.
- It’s not easy attracting these folks – DSs always look for quality of work and as long this is satisfied the rest should matter less. But yes not to be ignored.
- They come only from Big Brands- totally disagree, some of my best Data scientists were found at start-ups. Some even come with patents in their name and bring innovative solutions to the table.
- Compensation is a crowd puller – While Data Scientists are paid well, most of them have very clear priorities. They would be keen to know the role definition, growth plan, and company vision. If you are able to convince them on these aspects you can crack the deal on salary. But yes, you need to have decent budgets and make a lucrative offer.
- Data scientists are niche – this is a very common problem among all the skills these days. While you can find 1000s of DSs in the market, 10-20% would be relevant to you. So, search wisely.
- The normal interview process should do – Follow up to the previous statement, only a handful are relevant to your search. Hence the interview procedures must be modern and short. Hackathons and event based hiring solve the problem. More backdated your interview process, larger the chances of a drop-out. People like Belong, HackerRank, Hackerearth should be able to help you.
- DSs can work out of a garage- unless you are Apple or Facebook, starting up at your garage or the dorm, work on your Infra. Better the Infra, greater is the work that will be done. Give the Scientist the infra they ask & see the magic that is created. Happy employees create great customer success stories. Load them with gadgets & maintain a wonderful ambiance to sustain the ideal atmosphere for creativity.
Now that you have read all the myths, I’m sure you’d have the question on how I came up with these insights. I have been steadily hiring reporting folks in the early 2010 -11 and from then on there has been a lot of demand for statistics folks since a lot of statistical models were like products. It could be used for multiple analysis and upgraded from time to time & with the use of languages like python & R , machine learning algorithms were easily programmable.
Statistics & Tech is a deadly combination, if you get this right you can create magical solutions and give back some meaningful contribution to the company.
Some common traits of Data Scientists are as follows:
- A keen eye for detail – these folks are the ones who create insights from unstructured stuff. So unless you have an eye of an eagle, you cannot see through the challenges
- Tech know how- you need not be a rocket scientist, you still need to know advanced statistics and at least 2 programming languages or tools like R & Python. Hadoop, Scala etc.
- Communication is key – A Data scientist communicates with a wide audience, hence a very logical combination of articulation and flow is needed. A good Data scientist breaks down jargons into simple statements and helps others understand what S/He has done.
Some essential sources where you could find Quality Data Scientists
- Social Media – This is a very vague term in recent times, You can scourge facebook groups like (Data scientists, R , Python programmers, Analytics etc), go by the group which has a large number of members and the active posts
- Github, Stackoverflow will do the trick – This is like a playground to spot fine talent, sit with your hiring manager to overview codes written by them. This will eliminate the need for a deep technical interview.
- Events & Conferences – Grow out of your traditional job portals or even LinkedIn, there are plenty of events like Cypher, Pycon, Nasscom Data Analytics summits happening all through the year. Just go there and grab ‘em.
- Analytics portals/ Blogs – Lots of articles are being written online and you can spot folks who write the right thing and this will help spot talent that you are looking for.
Please do write to me to add any thoughts or questions on the topic.
Try deep learning using MATLAB