HackerRank opens Chakra interview agent to customers

HackerRank says Chakra completed more than 500,000 interviews during testing and is now generally available to customers. The release follows around six months in beta, according to the company. No regulator, legal status or statutory process is identified in the source material.

HackerRank describes Chakra as an AI agent that conducts interviews and evaluates candidates’ answers and reasoning. Its general availability is a product release, not a regulatory decision described by an authority. The source account does not state a date beyond Monday.

Testing preceded customer availability

HackerRank tested Chakra for around six months before announcing general availability. The company says it also used the product internally during that period. Snowflake, Snorkel and Capgemini were among the organisations that tried it, according to HackerRank.

The stated interview volume is more than 500,000. The source material does not explain how interviews were counted, what proportion involved each organisation or whether the figure represents completed sessions. It also gives no pricing or terms for customer access.

The interview includes AI assistance

A candidate using Chakra is given a task involving a real-world code repository, based on HackerRank’s description. The work happens in a canvas that includes an AI assistant. As a candidate proceeds, Chakra can use the work context to ask follow-up questions about choices and possible changes to the task.

HackerRank says the evaluation covers both the answer and how the candidate reached it. The company also says it wants to measure critical thinking, judgment and “AI fluency”. In its account, that last term concerns how a candidate presents a problem to AI, evaluates its output and steers it towards a solution.

HackerRank sets out its rationale

CEO Vivek Ravisankar told TechCrunch that employers may need to consider the thinking and judgment behind work products as AI makes it easier to produce an artefact. This is the company’s stated rationale for looking beyond final answers. The source does not provide an external assessment of whether Chakra can reliably measure those qualities.

Ravisankar also described a possible change to interview structure. He said a recruiter screen, a take-home assessment and an engineer follow-up interview could be combined into one Chakra interview. The account does not specify whether customers must adopt that format or how hiring teams should weigh the resulting report.

The product’s reported effect on suspicious activity is also presented by HackerRank. Ravisankar said flags were 70% to 80% lower than in comparable traditional HackerRank assessments, with variation attributed to factors such as geography. The source does not provide the underlying data, comparison method or a response from an independent party.

The supplied material identifies no regulator, investigation, legal finding or formal rule applying to Chakra. It also does not state how customers should disclose the AI agent’s role to candidates or how assessment records are handled. Those matters remain open in the source account, rather than established requirements. HackerRank’s stated next step is customer availability following the Monday announcement.