A Data Scientist builds models and experiments that change decisions — and their value depends far more on framing the right question than on the algorithm they choose.
Indicative Indian market ranges. Metro and specialist roles sit at the upper end.
The strongest signal is a candidate who has told a business "the data cannot answer this" and been right. Statistical rigour is easy to fake in interviews, so ask about a result they later discovered was wrong — honest scientists have one and will describe it plainly.
When you already have reliable reporting and specific prediction or causal questions that analysis cannot answer. Hiring a scientist into a business without clean data usually produces an expensive analyst.
For research-heavy or novel-method work, yes. For most commercial problems, applied experience and business judgement matter more, and PhD candidates sometimes over-engineer solvable problems.
Hiring for algorithm knowledge and discovering the person cannot frame a business problem or explain a result to a stakeholder. Test both explicitly.
Post the vacancy free and reach candidates who are actively looking. Most employers get their first applications within 24 hours.