HackerRank makes Chakra interviewer generally available

Candidates work with an embedded SI assistant while Chakra evaluates their approach, not just their answers. People retain the final hiring decision.

Candidates work with an embedded SI assistant while Chakra evaluates their approach, not just their answers. People retain the final hiring decision.
HackerRank is making its SI interviewer Chakra generally available to customers on October 5 after about six months in beta testing, TechCrunch reported. The developer-assessment and hiring platform is moving beyond tests focused largely on correct coding answers: Chakra observes candidates’ work and aims to assess the judgment behind their solutions.
That changes what candidates are asked to demonstrate. HackerRank wants to measure what it calls “AI fluency”: how well a candidate frames a problem for an assistant, evaluates its output and guides it toward a solution. Using an assistant is part of the interview, rather than something candidates must do secretly.
Candidates receive a task involving a real-world code repository—a collection of a project’s code—and work in a workspace with an embedded SI assistant. Chakra can use their ongoing work to ask why they chose one approach over another, or how they would change a solution if a new constraint were added.
After the interview, Chakra provides the hiring manager with scores and a detailed report. HackerRank co-founder and CEO Vivek Ravisankar told TechCrunch that one Chakra interview combines stages previously handled through a recruiter screening, a take-home assessment and a follow-up interview with an engineer.
Analysis
HackerRank says Chakra conducted more than 500,000 interviews during testing. Ravisankar said suspicious-activity flags were 70–80% lower than in comparable traditional HackerRank assessments, with the reduction varying by factors including geography and seniority.
Ravisankar attributed the lower flag rate to candidates having less incentive to use outside tools secretly when an assistant is already available. Fewer flags do not by themselves establish less cheating or more accurate assessment. How HackerRank matched the assessments for comparison, and whether independent testing would confirm the results, remain unclear.
Ravisankar said Chakra is designed to score candidates, not make the final hiring decision. In his description, the system applies criteria set by an employer, while human interviewers can focus on whether they want to work with a candidate and answer questions about the team and role.
Consistent scoring does not establish freedom from bias. TechCrunch cautions that automated hiring tools can inherit or amplify biases from their data, models and evaluation criteria.
Employers also face rules governing some uses of automated hiring tools. New York City requires employers using certain automated employment decision tools to obtain an independent bias audit and notify candidates before use. Ravisankar acknowledged that hiring is regulated and said HackerRank had built for compliance with such requirements.

The state wants third-party approval of safety measures before new models are developed. The request is not an issued court order.
1 min

Analysts say memory suppliers are prioritizing data centers, while phone makers cut investment in entry-level models.
3 min

The case concerns payments for articles and snippets in Search, News and Discover. A finding of abuse could carry a fine of up to 10% of a company’s turnover.
2 min