AIG (American International Group) Data Science Interview Questions and Answers (2026 Guide)

Quick answer: AIG's India Global Capability Center in Gurugram is a large, long-established insurance operations hub, roughly 4,000 people, running since 2008, built around actuarial modelling, claims technology and underwriting analytics rather than a stand-alone "data science" division. We could not find any India-specific candidate account of a Data Scientist or Data Analyst interview at AIG. What is documented on Glassdoor describes AIG's broader, largely US-flavoured process: for Data Scientist, a roughly 2-hour online test followed by a 1-hour technical phone interview; for Analyst-track roles, an HR screen, statistics questions, and sometimes a written SQL exam. Treat this as representative of AIG generally, not confirmed as the India loop specifically.

About AIG

AIG (American International Group) is a US-headquartered global insurer covering property, casualty, life and specialty lines in more than 80 countries. Its India Technology and Operations Centre, based in Gurugram's Unitech Cyber Park with additional offices in Mumbai, Bangalore, Hyderabad, Pune, Chennai and Ahmedabad, has run since 2008 and had grown to about 4,000 professionals, with AIG reportedly targeting 4,500 by 2027. The centre's own described focus is actuarial modelling, claims processing systems, underwriting analytics and the data infrastructure behind AIG's commercial and specialty insurance book, plus more recent AI work on claims severity assessment and fraud detection. (Business of GCC) We found this figure and history on a single GCC-tracking site with no second independent source repeating the exact numbers, so treat the headcount and 2027 target as reported rather than AIG-confirmed.

AIG's Interview Process

• Data Scientist (Glassdoor, location not specified): candidate reports describe a first round of largely conceptual machine-learning questions built around the candidate's own resume projects, including comparing classifiers and explaining the difference between Lasso and Ridge regression, followed by a roughly 2-hour online test with at least 3 coding problems, then a 1-hour phone interview with two interviewers focused on machine-learning concepts and basic applied math such as probability, derivatives and expectation. (Glassdoor)

• Analyst-track roles (Glassdoor): the process commonly opens with a recruiter or HR call of about 30 minutes on your resume and fit, followed by technical questions on statistics (R-squared versus adjusted R-squared, p-values, classification and clustering algorithms, how to validate a model). One reported path ran across two days: a written SQL exam on day one, then a 45 to 60 minute technical round and an HR round on day two. Glassdoor's aggregate for Analyst-level roles: 2.6 out of 5 difficulty, 58% positive experience, and an average 33 days to hire. (Glassdoor; Glassdoor)

What Candidates Report

Beyond process mechanics, Glassdoor's location-filtered reviews give a useful, if mixed, signal on AIG's Bangalore office specifically: a 4.5 out of 5 rating from 261 reviews, but only 62% of reviewers would recommend working there to a friend, against 83% company-wide. (Glassdoor) Separately, one review under AIG's Data Scientist reviews describes a team where roughly one data scientist resigned per week out of about 50 in the department, alongside frequent reorganisations, about twice a year, and limited growth for staff without senior-management connections. (Glassdoor) We are flagging this as a single reviewer's account, not confirmed as India-specific or as representative of the team today.

What We Could Not Verify

We could not find one India-specific, first-person account of a Data Scientist or Data Analyst interview at AIG: Naukri Code360's AIG interview bundle blocked automated access, AmbitionBox returned nothing usable in search, and GeeksforGeeks had no AIG data-role write-up. Every Glassdoor page above also returned an access error on direct fetch, so the process detail comes from search-indexed summaries rather than pages we read in full, and none of them state whether the reviewer sat in India, the US or elsewhere. We also could not confirm that AIG's Gurugram centre runs a distinct "Data Scientist" job family separate from its actuarial and underwriting analytics teams; its own public description centres on those functions, which is plausible context but not a confirmed org-chart fact. Two prep-guide sites, InterviewQuery and Dataford, list AIG data-role "guides," but their content read as generic keyword summaries rather than sourced candidate intelligence, so we have not repeated specific claims from them here.

How to Prepare

• Know your own projects cold: the clearest reported pattern is being asked to defend model choices, like Lasso versus Ridge, from your own resume rather than from a textbook.

• Revise core statistics: R-squared versus adjusted R-squared, p-values, and the practical difference between classification and clustering come up repeatedly in Analyst-track reviews.

• Prepare for a written SQL component that may be separate from your technical interview, not folded into the same session.

• Think in insurance terms: given the Gurugram centre's actuarial, claims and underwriting focus, be ready to reason about risk, fraud detection or pricing use cases, even though we could not confirm this is asked directly.