HackerRank is making Chakra, its AI coding interviewer, generally available after roughly six months in beta and more than 500,000 interviews. Early testers included Snowflake, Snorkel and Capgemini, according to the company.
Candidates work on a realistic code repository with an AI assistant rather than solving an isolated coding puzzle. Chakra watches the process, asks follow-up questions about choices or changed constraints, and evaluates problem framing, judgment and the ability to assess AI output. HackerRank says the product can combine a recruiter screen, take-home assessment and engineer interview into one session.
The company reports that suspicious-activity flags were 70% to 80% lower than in comparable traditional assessments, with variation by geography and seniority. Its explanation is that permitted AI use reduces the incentive to hide outside assistance. That is a company result, not an independent finding about cheating or validity.
Chakra scores candidates but leaves the final hiring decision to people. Consistent rubrics can reduce some interviewer variability, yet models and criteria can still reproduce bias. Employers must validate results across demographic groups and follow local rules; New York City, for example, requires certain automated hiring tools to receive an independent bias audit and requires notice to candidates.