HackerRank Opens Its AI Interviewer to Customers, Allowing AI in Coding Tasks

Developers interviewing through HackerRank’s Chakra can use an AI assistant—and face questions from an AI interviewer about their choices. HackerRank made the product generally available to customers on October 5 after roughly six months in beta. The system scores candidates’ work, while humans retain the final hiring decision, according to the company.

HackerRank built its business around coding challenges that tested whether developers could solve problems correctly. Co-founder and CEO Vivek Ravisankar told TechCrunch that AI has made that approach less useful: producing a finished answer no longer tells an employer enough about the thinking behind it.

Chakra is intended to assess critical thinking, judgment and what HackerRank calls “AI fluency”: framing a problem for AI, evaluating its output and steering it toward a solution. The company says it conducted more than 500,000 interviews during testing, with Snowflake, Snorkel and Capgemini among the participants.

Candidates receive a task involving a real-world code repository—a collection of a project’s code—and work in a canvas with an AI assistant. Chakra can ask why they chose one approach over another, or how their solution would change under a new constraint. Those questions draw on what the candidate is doing.

Ravisankar says a single Chakra interview combines steps that previously required a recruiter screen, a take-home assessment and a follow-up interview with an engineer. Hiring managers receive a detailed report afterward.

HackerRank’s July explanation of Chakra’s scoring describes three separate AI agents. One works with the recruiter to turn a job description into an interview plan with explicit expectations. Another conducts the live, voice-based interview. A third reads the transcript afterward and scores each expectation against that plan.

The company says every score traces to a specific moment in the transcript. Criteria are rated as met, partially met, not met or not assessed, then normalized to a 0–5 scale for the recruiter’s report. If no relevant transcript evidence exists, the criterion is marked “Not Assessed” rather than assigned a guessed rating.

Ravisankar argues that consistently applying an employer’s criteria can make AI less biased than people. Consistency alone does not establish fairness: automated hiring tools can inherit bias from their models, data and criteria. New York City requires independent bias audits and candidate notice for certain automated employment decision tools; Ravisankar says HackerRank has built for such requirements.