Elimika
Intelligent Exam Integrity and Automated Grading

Exams That Mean Something

Elimika 365 gives institutions the infrastructure to conduct online examinations that are fair, supervised, and accurate — with real-time integrity monitoring and AI-assisted marking that eliminates the failure points that have made online assessment unreliable.

A product by Martial Jenga Labs.

Elimika 365

What We Build

Two Phases. One Complete Assessment.

Elimika 365 operates in two connected phases — examination and grading — each engineered to remove the failure points that have made online assessment unreliable.

Phase One

The Examination Phase

Locked Examination Environment

The browser locks down the moment a student enters a session. Navigation away from the exam is blocked, and any attempt to do so is recorded as a flagged event.

Continuous AI Monitoring

The student's camera and microphone are active throughout. The AI analyses the full scope of what is visible — the environment, the student's presence and attention, and patterns of behaviour across the entire session.

Personalised Integrity Scoring

Analysis is not applied against a generic standard. The system builds an understanding of each individual student, producing an integrity score that is personalised and significantly more accurate than tools that measure every student against the same fixed benchmark.

Real-Time Visibility

The assigned lecturer has live visibility of flagged events as they occur. When the session ends, a full integrity report is saved alongside a recorded video, with each flag linked to a precise timestamp.

Phase Two

The Grading Phase

Fully Automatic Marking for Elimika Learn Exams

Exams originating from Elimika Learn are marked fully automatically. Because Elimika Learn generated the questions and holds the full context of the lesson plan and learning objectives, the grading model has everything it needs without any lecturer setup.

Rubric-Guided Marking for Custom Question Banks

For exams from an institution's own question bank, the lecturer submits their marking rubric and refines it through a guided interface until the model's approach meets their standard. The model then marks the full paper to those agreed criteria.

Collective Intelligence

The model learns from every lecturer interaction. Each correction, refinement, and rubric submitted improves marking intelligence over time, and those improvements are reflected across the platform.

AI-Use Detection

Every answer is evaluated for indicators that the response was generated by an AI tool rather than written by the student. Suspicious results are surfaced to the administrator. Clean answers are finalised without administrative involvement.

The Problem

Why Online Examinations Are Failing Institutions

Three compounding challenges that no existing tool has fully addressed — and one that compounds them all.

Opportunity01

Cheating Is Always Possible

When a student sits an exam outside a supervised physical space, the opportunity to seek assistance — from another person, a device, or the internet — is always present. No amount of policy language changes the structural reality that online examination, without active supervision, is unverifiable.

Scale02

Manual Review Does Not Scale

Manually reviewing footage, marking papers, and auditing results for every student in every exam is not operationally viable for any institution of meaningful size. Every hour an administrator spends reviewing routine, clean sessions is an hour wasted. Existing tools produce too much noise to act on.

Fairness03

Inconsistency Undermines Integrity

Without consistent, objective standards applied uniformly across all students, integrity decisions are vulnerable to challenge, bias, and inconsistency. The same script marked by two different lecturers can yield two different scores. The same behaviour flagged in one session may go unnoticed in another.

Compounding the Problem

Marking Is Broken Too

Unconscious Bias

Lecturers may unknowingly mark students they know differently. Results are never truly objective when the marker knows the candidate.

Missed and Miscalculated Marks

Manual marking at scale is error-prone. Marks get skipped, totals go wrong, and students pay the price for administrative failure.

No Audit Trail

When disputes arise after an exam, there is no recorded evidence to reference. Institutions are left exposed and unable to defend their results.

The Solution

How Elimika 365 Works

Two connected pipelines — examination and grading — each engineered to remove human failure points at every stage.

Examination Phase

Supervised, Secured, and Scored in Real Time

From the moment the exam starts to the moment the session ends, Elimika 365 monitors every student with precision.

4 Steps
01

Secure Environment Setup

The browser locks down the moment the student enters the session. Navigation is blocked. Camera and microphone activate. The student's answers transmit in real time and are saved securely throughout.

02

Identity Verification and Baseline

Each student is verified before the exam begins. A personal behavioural baseline is established during calibration so the system knows who it is watching — and flags deviations from that baseline with precision, not against a generic standard.

03

Continuous AI Integrity Monitoring

The AI analyses the full scope of what is visible — the physical environment, presence, attention, and behavioural patterns throughout. It produces a single consolidated integrity score. Findings are meaningful, well-evidenced, and sparse enough to be actionable.

04

Timestamped Report and Full Audit Trail

When the session ends, a full integrity report is saved alongside a recorded video. Each flag is linked to a precise timestamp. Direct engagement with footage happens only when a finding is sufficiently unusual, or when a student formally requests review.

Grading Phase

Consistent, Traceable, and Bias-Free Marking

The grading pipeline begins automatically when a session ends. No manual intervention on clean submissions.

4 Steps
01

Automatic Marking for Elimika Learn Exams

Exams generated by Elimika Learn are marked fully automatically. The grading model has full context of the lesson plan, learning objectives, and expected outcomes — no lecturer input required.

02

Rubric Calibration for Custom Exams

For institutions using their own question banks, the lecturer submits their marking rubric and refines it through a guided interface. The model marks a sample, the lecturer corrects it, and the rubric is finalised. The model then marks the full paper to those agreed standards.

03

AI-Generated Response Detection

Every answer is evaluated for indicators that the response was generated by an AI tool rather than written by the student. Suspicious results are surfaced to the administrator for a decision. Clean answers are finalised without any administrative involvement.

04

Results Returned and Credential Issued

Results are produced, recorded against the learner's profile, and where applicable a verified credential is issued. The entire pipeline — supervision, marking, detection, certification — runs without manual intervention on clean sessions.

Who It's For

Built for Everyone in the Examination Process

Whether you run an institution, teach a class, or sit the exam — Elimika 365 was built with your specific stakes in mind.

Take Back Control of Online Examinations

Academic institutions carry the responsibility of certifying knowledge fairly and accurately. Elimika 365 gives you the infrastructure to do that online — with real-time supervision that deters and detects misconduct, AI-assisted marking that eliminates inconsistency, and a full audit trail that protects your institution when results are challenged.

  • Real-time anomaly detection across all active exam sessions
  • AI marking that is consistent, traceable, and bias-free
  • Full session logs and audit trail for every exam
  • Scalable from single courses to institution-wide deployment

Mark Less. Trust More. Stay in Control.

Marking hundreds of scripts manually is exhausting, error-prone, and time-consuming. Elimika 365 handles the heavy lifting — AI processes each script with consistent standards. You review outcomes rather than grind through answers. During exams, the system monitors students on your behalf and surfaces only the alerts that require your attention.

  • AI-assisted marking reduces your workload significantly
  • Live supervision alerts — you intervene only when needed
  • Review and approve marking decisions — full control stays with you
  • No more disputed marks or missing scores

An Exam You Can Trust

For students sitting honestly, Elimika 365 is a guarantee that their effort is assessed fairly. Marks are not missed. Standards are applied consistently. If a specific incident is flagged and a student believes it is incorrect, they can formally request that the specific moment be reviewed against the timestamped recording.

  • Consistent marking standards applied to every student equally
  • No missed marks or miscalculated totals
  • Formal review mechanism tied to timestamped session footage
  • Integrity score built on individual behavioural baseline, not generic rules
Connected Ecosystem

The Integration

One Pipeline from First Lesson to Final Credential

Elimika 365 is one half of a connected ecosystem. When a learner is ready to be assessed, the exam flows directly from Elimika Learn — no configuration, no duplication, no gap between learning and certification.

Learn about Elimika Learn
1
Learn

Knowledge Built in Elimika Learn

The learner builds knowledge through Elimika Learn — personalised lessons, adaptive content, and continuous progression tracked in a living learner profile.

2
Generate

Exam Generated Automatically

When the learner is ready to demonstrate competency, Elimika Learn generates the exam. Because it holds the full context of the lesson plan and learning objectives, no lecturer setup is required.

3
Supervise

Examina 365 Takes Over

Elimika 365 administers the exam under full integrity monitoring. The AI supervises the session, flags anomalies in real time, and produces a consolidated integrity score.

4
Mark

Graded Without Manual Intervention

Scripts are marked automatically using the learning objectives and expected outcomes from Elimika Learn. No lecturer input required on clean sessions.

5
Certify

Verified Credential Issued

Results are returned, recorded against the learner's profile, and a verified credential is issued. The qualification carries the full weight of both platforms behind it.

Get In Touch

See Elimika 365 in Action

Whether you are an institution exploring online examination infrastructure or an educator looking to eliminate marking burden — let us walk you through it.