HackerEarth
HireAIScore rates HackerEarth 46 out of 100 (grade F, Substantial gaps) under rubric v1.0, last reviewed September 25, 2026. HackerEarth ranks 39 of 82 vendors rated in Assessment AI, against a category average of 46. That score is drawn from 39 evidence items across 7 rubric criteria for HackerEarth — 34 with a cited source, 5 recording that nothing was located.
Score breakdown
HackerEarth, criterion by criterion
Each rubric criterion is scored out of 100 on the cited evidence, then weighted into the total.
Weighted contribution shown to the right of each bar.
Article 11 Technical Documentation
Weight 20% · category avg 50
55
+11.0
Bias Audit Transparency
Weight 18% · category avg 41
35
+6.3
FRIA Support
Weight 15% · category avg 31
32
+4.8
Data Governance Disclosure
Weight 15% · category avg 53
52
+7.8
Human Oversight Design
Weight 12% · category avg 55
56
+6.7
Post-Market Monitoring
Weight 12% · category avg 38
42
+5.0
Customer Documentation
Weight 8% · category avg 58
57
+4.6
Total weighted score
46.2(rounded to 46 for display)
Category avg is the mean raw score on that criterion across the 82 Assessment AI vendors in scope of this rubric, this one included.
Strongest criterion
Customer Documentation
57raw / 100
Contributes 4.6 to the total. 57 against a 58 category average.
Weakest criterion
FRIA Support
32raw / 100
Contributes 4.8 to the total. 32 against a 31 category average.
Cited evidence
The evidence behind the score
Every score is backed by at least one cited piece of evidence. Open a criterion to see what was found, when, and where.
Download diligence recordEvidence ledger
- Items
- 39
- Documentation
- 30
- Public statement
- 4
- Absence
- 5
- With a source URL
- 34 of 39
- Source hostnames
- 2
- Fewest items
- 3
- FRIA Support
These figures measure how thoroughly HackerEarth was reviewed, not how HackerEarth performed — an absence row, recording that nothing was located, is counted like any other item.
Article 11 Technical Documentation
6 items55
- DocumentationCaptured Sep 25, 2026
The Trust Center's 'AI Posture' section says candidate data is never used to train models, hiring decisions stay with the customer, and nothing scores emotion, faces or sentiment. It says every score traces to rubric-mapped evidence, lists ISO 42001:2023 as certified, and offers an 'Artificial Intelligence Security Policy' only on request.
- DocumentationCaptured Sep 25, 2026
The OnScreen product page describes a deterministic evaluation framework that probes the same skills with the same rubric for every candidate, and reports that include competency-level scores, a transcript highlighting the exchanges behind each score, the candidate's code, and integrity signals.
- DocumentationCaptured Sep 25, 2026
A help-center system description says OnScreen is a voice-mode AI interviewer with template-defined skills that produces an overall score out of 100 plus per-skill scores, and that its toggleable integrity controls (AI Likeness Detection, candidate video, Smart Browser, ID verification) report as an integrity index.
- DocumentationCaptured Sep 25, 2026
This article documents that AI evaluation of subjective answers uses a proprietary model to compare the candidate's text with a recruiter-supplied base answer, and warns not to use it for numerical or non-English answers.
https://help.hackerearth.com/evaluation-of-subjective-questions
- DocumentationCaptured Sep 25, 2026
The help center documents two plagiarism-detection methods: line-by-line comparison between candidates, flagged above an 85% match, and an AI logic-based internet check reported as a Medium or High 'Plagiarism Likeliness'.
https://help.hackerearth.com/hc/en-us/articles/360003777793-plagiarism-insights
- AbsenceCaptured Sep 25, 2026
The trust center, OnScreen page, help center and blog contain no model or system cards, validation or accuracy studies, technical documentation pack, or instructions for use for OnScreen, AI answer evaluation or the Integrity Index. The claimed ISO 42001 certificate is also missing from the Trust Center's document list.
Bias Audit Transparency
5 items35
- Public statementCaptured Sep 25, 2026
The Trust Center claims 'Periodic adverse-impact analysis, with methodology designed by industrial-organizational psychologists' but names no auditor, dates or results, and no bias audit appears in its document list.
- Public statementCaptured Sep 25, 2026
The OnScreen launch post claims one enterprise customer screened more than 2,000 candidates in a weekend 'with complete consistency and zero interviewer bias', with no published fairness testing behind the claim.
- DocumentationCaptured Sep 25, 2026
OnScreen analytics show customers score distributions by integrity band, status breakdowns and funnel metrics, but no demographic or adverse-impact breakdowns.
https://help.hackerearth.com/interview-analytics-in-onscreen
- Public statementCaptured Sep 25, 2026
HackerEarth's June 2026 buyer guide tells hiring teams to demand each vendor's third-party bias-audit methodology and demographic breakdown, which HackerEarth itself does not publish.
- AbsenceCaptured Sep 25, 2026
Web searches, the Trust Center, help center and blog turned up no NYC LL 144 bias-audit summary, no independent audit (e.g., BABL AI, DCI, Warden AI, Holistic AI) and no impact ratios for OnScreen scores, AI answer evaluation or proctoring/Integrity Index flags. trust.warden-ai.com/hackerearth returns 404.
FRIA Support
3 items32
- DocumentationCaptured Sep 25, 2026
The Trust Center lists a DPA, 31 security and AI policies, four ISO certificates, penetration-test reports and a SOC 2 Type II report, all behind 'Request Access' (NDA), which gives deployers a partial diligence package.
- Public statementCaptured Sep 25, 2026
HackerEarth's overview of AI-interview regulation (NYC LL 144, EU AI Act, Illinois AIVIA) is generic and sends deployers to legal counsel instead of providing HackerEarth-specific templates.
- AbsenceCaptured Sep 25, 2026
The Trust Center, OnScreen page, help center (including the GDPR FAQs) and blog contain no EU AI Act Article 27 FRIA template, Article 26 deployer-obligation guidance or instructions for use.
Data Governance Disclosure
6 items52
- DocumentationCaptured Sep 25, 2026
The Trust Center lists ISO 27001:2022, 27017, 27701 and 27018 certificates (valid through Mar 2029) and DPF certification, and commits never to train models on candidate data. It sets a three-year maximum retention that customers can shorten, 30-day erasure, 72-hour breach notice, and names subprocessors including Tavus as an 'AI service provider supporting Proctoring'.
- DocumentationCaptured Sep 25, 2026
The privacy policy (updated September 15, 2025) names HackerEarth Inc. as controller or processor, confirms DPF certification and says personal data is not subject to GDPR automated decision-making, but never mentions webcam snapshots, audio/video recordings, government-ID checks or face matching.
- DocumentationCaptured Sep 25, 2026
The candidate FAQ says data is kept until the candidate or test organizer asks for deletion, which contradicts the Trust Center's three-year maximum retention.
https://help.hackerearth.com/hc/en-us/articles/360005045413-for-how-long-will-my-data-be-stored-
- DocumentationCaptured Sep 25, 2026
Candidates cannot delete their own webcam snapshots or video; the test organizer must ask HackerEarth, which deletes them 30 days after receiving the request.
- DocumentationCaptured Sep 25, 2026
The security policy (updated September 15, 2025) calls its list of third-party processors (SendGrid, Plivo, Twilio, AWS, Landbot) 'exhaustive', but the list does not match the Trust Center's subprocessor list.
- DocumentationCaptured Sep 25, 2026
A December 2023 product post says AI evaluation of subjective answers 'uses ChatGPT and HackerEarth's proprietary AI models', a third-party model dependency that neither published subprocessor list mentions.
Human Oversight Design
8 items56
- DocumentationCaptured Sep 25, 2026
The OnScreen page says it never makes the hiring decision and only gives an evidence-backed recommendation, which hiring managers can trace through transcript exchanges, the candidate's code and the rubric.
- DocumentationCaptured Sep 25, 2026
An OnScreen report shows the recruiter overall and per-skill scores, the integrity index, the AI-likeness result and proctoring details, with a downloadable transcript. The only documented actions are archiving the candidate or resetting the interview, with no way to edit or override the AI score.
https://help.hackerearth.com/downloading-candidate-data-in-onscreen
- DocumentationCaptured Sep 25, 2026
Diversity & Inclusion settings let admins hide candidate PII from tests and interviews, using aliases in reports and interview screens and disabling snapshot capture.
https://help.hackerearth.com/hc/en-us/articles/900001160763-personally-identifiable-information-pii-
- DocumentationCaptured Sep 25, 2026
Recruiters can grade subjective answers manually or by keyword instead of with AI.
https://help.hackerearth.com/evaluation-of-subjective-questions
- DocumentationCaptured Sep 25, 2026
An AI-computed Integrity Index rolls proctoring violations up into Acceptable, Marginally Acceptable or Unacceptable, and marginal cases are sent for further review rather than disqualified automatically.
- DocumentationCaptured Sep 25, 2026
Smart Browser automatically removes candidates from the test if they do not return within 10 seconds of a warning for leaving the test screen, with no human review.
https://help.hackerearth.com/advanced-proctoring-settings-smart-browser
- DocumentationCaptured Sep 25, 2026
Recruiter-configurable proctoring includes a 'Log out on leaving the test environment' setting (5 warnings by default, and closing OS notifications counts as leaving) and webcam snapshots that stop candidates who deny camera access from taking the test.
https://help.hackerearth.com/hc/en-us/articles/360002987633-proctoring-settings
- DocumentationCaptured Sep 25, 2026
The proctoring feature page markets AI eye-movement tracking, face recognition and machine-learning detection of 'suspicious gestures' (for example, looking to the side too often), with flagged clips time-stamped for recruiter review.
Post-Market Monitoring
5 items42
- DocumentationCaptured Sep 25, 2026
Public monthly release notes from 2020 through August 2026 document product, proctoring and scoring changes, but the Root Cause Analysis section contains only one RCA, for a 2019 email-delivery failure.
- DocumentationCaptured Sep 25, 2026
The August 2026 release notes disclose that for events published after August 11, 2026, audio-proctoring results now reduce the Integrity Index (a major deduction for high_risk, a minor one for needs_review), which works as a de facto log of scoring changes.
- DocumentationCaptured Sep 25, 2026
HackerEarth runs a bug bounty program with scoped domains, severity-based rewards and a 24-48 hour acknowledgment target, with reports sent to support@hackerearth.com.
- DocumentationCaptured Sep 25, 2026
The security policy sends incident reports to the general support@hackerearth.com inbox, says customer-impacting incidents are reported based on severity, and states that third parties run penetration tests annually.
- AbsenceCaptured Sep 25, 2026
HackerEarth has no public AI performance or fairness monitoring dashboard, no AI-incident reporting channel, no security.txt (404) and no dedicated security@ contact.
Customer Documentation
6 items57
- DocumentationCaptured Sep 25, 2026
A dedicated OnScreen help section covers interview templates, invites, report downloads, analytics and proctoring settings.
- DocumentationCaptured Sep 25, 2026
GDPR FAQs for candidates and recruiters cover data deletion, retention, audit logs, access and third-party sharing.
https://help.hackerearth.com/frequently-asked-questions-faqs
- DocumentationCaptured Sep 25, 2026
The Trust Center brings together certifications, AI posture, security controls, subprocessors, and a DPA and policies available on request.
- DocumentationCaptured Sep 25, 2026
The OnScreen page's FAQ answers buyer questions on facial analysis, identity verification, who makes the hiring decision, ATS integration and pricing, citing NYC LL 144, the Illinois AI Video Interview Act and the EU AI Act.
- DocumentationCaptured Sep 25, 2026
The FaceCode page explains what feeds its AI interview summaries (code, transcript, interviewer scores and behavioral signals) and claims an EEOC- and OFCCP-compliant design with 'Compliance documentation available on request.'
- AbsenceCaptured Sep 25, 2026
The trust center, product pages, help center and blog contain no HackerEarth-specific NYC LL 144 candidate-notice guidance, Illinois AIVIA consent workflow, or Colorado or EU AI Act deployer guide; regulatory content is limited to marketing FAQs and generic blog posts.
Editorial notes
Company overview
Sachin Gupta and Vivek Prakash founded HackerEarth in November 2012. The legal entity is HackerEarth Inc. Its privacy policy gives a Sunnyvale, California address (older profiles say San Francisco), the company also has a large Bengaluru base, and it named Vikas Aditya CEO in July 2024. Its Assessments product auto-scores coding tests with test suites and grades written answers with AI. Both run under AI proctoring (Smart Browser lockdown, webcam snapshots, gaze tracking, face recognition and audio monitoring), which feeds a single 'Integrity Index'. FaceCode runs live panel coding interviews with AI-compiled summaries. OnScreen launched on April 14, 2026. It is a lifelike AI video-avatar interviewer with KYC-grade ID checks and face matching, and it scores each candidate 0-100 overall and per skill against a fixed rubric. Published pricing starts at $99/month for assessments and $129/month (10 interviews) for OnScreen.
Regulatory exposure
HackerEarth's scored coding tests, AI-graded interviews and proctoring (which can end a test session) all evaluate job candidates. That puts it in the EU AI Act's Annex III employment high-risk category. The EU's Digital Omnibus deal moves those obligations to December 2027, but Article 50 duties to tell people they are talking to an AI apply from August 2026, which matters for an avatar interviewer. When employers rely on OnScreen's 0-100 scores or filter candidates by Integrity Index, those tools can count as automated employment decision tools under NYC Local Law 144, and HackerEarth has published no bias audit. AI analysis of recorded interview video can bring in the Illinois AI Video Interview Act. Matching faces against government IDs raises Illinois BIPA questions that the privacy policy never addresses. The Illinois Human Rights Act amendment, in force since January 1, 2026, requires notice when AI is used. Colorado's SB 26-189 (effective January 1, 2027) will require developer documentation on intended uses, training-data categories, limitations and human review. HackerEarth's refusal to score faces or emotions helps. However, gaze tracking, machine-learning 'behavioral pattern detection' and the automatic removal of candidates who stay off the test screen for 10 seconds after a warning remain live risks of disparate impact and of failing to accommodate disabilities.
Path to a higher score
HackerEarth should commission and publish an independent LL 144-style bias audit covering OnScreen scores and proctoring/Integrity Index outcomes, with impact ratios by sex, race/ethnicity and their intersections. It should also publish the results of the 'periodic adverse-impact analysis' its trust center mentions. It should support the ISO 42001 claim with the certificate, the issuing body and the scope, and resolve its conflicting SOC 2 ('audited' vs 'underway'), retention and subprocessor statements. It should publish system cards and instructions for use for OnScreen, AI answer evaluation and the Integrity Index, including face-match and gaze-detection error rates by demographic group. It should add recordings, ID documents and biometric data to the privacy policy. It should replace automatic proctoring removals with human review and document how scores can be overridden and how candidates can get accommodations. It should give customers deployer kits: NYC candidate-notice templates, Illinois AIVIA and BIPA consent flows, Colorado developer disclosures and EU AI Act FRIA templates. Finally, it should put real components and incident history on its dormant status page.
Regulatory frame
What applies to assessment AI.
Assessment is the most-regulated surface in AI hiring. NYC LL 144, EEOC and OFCCP frameworks, Illinois HB 3773, and EU AI Act Annex III §4 all apply, with overlap on bias-audit obligations.
Compare
HackerEarth against its nearest-scoring peers.
The closest scores in Assessment AI, side by side.
- Side-by-sideHackerEarth vs 365Talents46F
- Side-by-sideHackerEarth vs Korn Ferry46F
- Side-by-sideHackerEarth vs TalentLens46F
- Side-by-sideHackerEarth vs Talkpush46F
Ranking
Others rated in Assessment AI
Ranked 39 of 82 by weighted total under rubric v1.0. The ordering is arithmetic on the rubric and carries no view on which tool suits a given hiring process.
- #36Knockri47F
- #37TestGorilla47F
- #38365Talents46F
- #39HackerEarthThis profile46F
- #40Korn Ferry46F
- #41TalentLens46F
Conflicts of interest
No vendor pays for placement, scoring, or removal. Casework - the consulting firm that operates this directory - provides paid services to some vendors. Any active or recent (within 24 months) commercial relationship is disclosed on the affected vendor profile and the review is reassigned to an independent reviewer. See the full policy on About.
Casework has no commercial relationship with this vendor.