Related Posts
I have following offers in hand and my last working day is 14th June. I am 3 years experienced in Automation Testing. 1. Accenture - 9 LPA (7.5 fixed) 2. Mindtree - 9.5 LPA(fixed) 3. Mindfire Solutions - 10.16 (Fixed) 4. TIAA - 12 Fixed 5. EPAM Systems - 13 Fixed + 1.5 Joining bonus How should I plan to decline to companies and how should I try to ask them to raise my salary? Infosys Newco Tata Consultancy Wipro Mindtree EPAM Systems
How will I get onsite opportunity in Accenture?
Thoughts on Havas Health?
Additional Posts in Data & Analytics Consultants
New to Fishbowl?
unlock all discussions on Fishbowl.
I usually struggle with people not having soft skills. Great data scientists are great story tellers. Technical acumen is a necessary of course, but don't look down on ppt.
We need to explain analyses clearly and concisely to non data scientists. We don't work at Google and our clients typically don't have the same technical acumen as those sorts of folks. Sometimes the best model is the one you can explain.
Also there is a big push for model interpretability. It is increasingly important to understand how you can utilize libraries or develop algorithms to approximate how a model makes predictions.
I see lots of resumes with tons of R and Python packages listed but very little understanding of modeling and interpretation during interviews. You need to apply some business intuition and think about an approach that makes sense to solve the problem at hand. R and Python are great, but they are tools to computationally solve the problem you have formulated.
Humility.
A lot of data scientists come in talking about how awesome approach A, B, and C are without realizing professional data concerns.
Data is often horribly fractured, sparse, or biased unlike competition or classroom data. They lack experience in dealing with highly imperfect data and the understanding of the criticality in proper data cleaning, and the implications of the decisions made during that phase.
The number of do-overs that occur due to recklessness is horrid.
I agree about a lack of good storytelling skills. I'd also add the ability to ask good questions.
In contests/classrooms, the question you're trying to answer is literally given to you. A lot of clients don't really know what they want, or what they can find out. It's our job to help them figure out what questions to ask of their data. That's difficult to teach people to do. It requires, for lack of a better term, good taste.
I struggle to find strong technical talent that I can put in front of an executive to articulate the problem and solution and tell a good story. I’m starting to favor candidates with soft skills (with a training commitment/budget) over deeply skilled technical candidates, and it’s working slowly.