Accenture's data scientist pipeline most consistently opens with a recruiter or HR screen. It checks educational background, work history, and fit, and one candidate was also asked about salary expectations.
A few candidates described extra steps around this stage. One completed SQL, Python, and a basic English assessment before the final technical interview. Another encountered a proctoring portal that asks you to read sentences aloud on camera so it can map your lip movements before the interview began. A campus candidate started with an hour-long online assessment on machine learning algorithms that gated entry to the interview rounds.
Live technical rounds lean on practical SQL, with joins and window functions coming up for multiple candidates. Examples include ranking users by transaction volume or finding users whose transaction counts exceed the average. Python questions range from basic concepts like decorators to resume-grounded coding, such as importing a library and fitting a tree-based model. One candidate also faced statistical coding on simulations and confidence intervals. Another was asked how to handle unrealistic or inaccurate data without losing useful records. Case-style questions tend to be verbal: one candidate was asked to frame a machine learning problem and pick a model, in a round cut from an hour to 30 minutes.
Expect detailed questions on your own projects. Interviewers asked candidates to explain their contributions and the algorithms listed in their CV. AI work came up repeatedly, from basic GenAI questions to LLMs, AI agents, and scaling challenges in a Senior Manager round.
The process usually closes with a managerial or HR round focused on culture fit, handling pressure, and difficult situations. In one case, a candidate with strong technical answers was still rejected over stated communication concerns. Practicing clear, structured explanations of your reasoning is worth the extra effort.
Synthesized from 14 candidate reports by our editorial team.
