Documentation

Candidate Ranking

Upload resumes and review AI evaluation results

The job page is the central workspace for resume upload, AI ranking, candidate selection, status updates, comparison, and export.

Upload Resumes

Veritik supports:

  • PDF (.pdf)
  • Microsoft Word Open XML (.docx)

You can select multiple files in one batch. Veritik validates the file type and signature before parsing it, processes up to three files concurrently, and reports Queued, Screening, Retrying, Completed, or Failed for each file independently.

Keep the browser open until the batch finishes. Every selected file receives a private upload-attempt identifier that is reused for automatic retries. Temporary service responses (429, 502, 503, and 504) receive up to two retries. After an ambiguous network failure, Veritik checks the server-side processing status and recovers an already saved candidate instead of creating a duplicate. The result list displays Checking saved result or Waiting for existing process while recovery is running.

Screening remains request-bound. If an attempt is still processing after the recovery window, keep the page open and refresh the job before starting a separate upload. Selecting the same file again in a later, independent file-picker session intentionally starts a new attempt.

Legacy .doc files are not supported. Convert them to PDF or DOCX before upload.

PDFs must contain a readable text layer. Scanned or image-only PDFs cannot be scored without OCR and are reported as requiring OCR. Veritik does not continue to AI scoring when text extraction is empty or unreadable.

For supported datasets and normal service conditions, the experience target is processing up to 50 CVs in approximately 3 minutes. Actual duration varies with file size, document complexity, network conditions, and AI service availability.

AI Processing Flow

For each valid resume, Veritik:

  1. Extracts resume text.
  2. Identifies candidate profile information.
  3. Evaluates skills, experience, and education.
  4. Applies the job's AI rubric.
  5. Produces category scores, criterion-level CV evidence, a summary, and an AI Recommendation.
  6. Adds the candidate to the job ranking with Recruitment Status New.

The AI produces the Skills, Experience, and Education category scores. Veritik then calculates overall_score on the server using the category weights selected when the job was created. For example, with 50% Skills, 35% Experience, and 15% Education:

overall_score = skills_score × 0.50
              + experience_score × 0.35
              + education_score × 0.15

The result is rounded to two decimal places. This keeps the recruiter's priorities authoritative instead of allowing the AI to choose its own final weighting. Legacy jobs without saved category weights use 40/40/20.

For newly screened candidates, the assessment also records one evidence item for every required skill, preferred skill, experience requirement, and education requirement. Evidence is reviewed in Candidate Profile rather than expanding every ranking card.

Candidate Results

Candidate cards show important screening information such as:

  • Candidate name and extracted details.
  • Overall AI score.
  • Evaluation breakdown.
  • AI Recommendation.
  • Recruitment Status.

Candidates can be sorted by score and filtered using available screening controls.

Score Scale and Bands

Overall, skills, experience, and education scores are percentages from 0 to 100. A score of 8 means 8%, not 8/10. Decimal percentages such as 85.5 are valid, and 0 must be shown as 0% rather than as missing.

The current ranking filters use these score bands:

  • Recommended: 85 or higher.
  • Consider: 70 or higher and below 85.
  • Not Recommended: below 70.

These bands help prioritize review but do not update Recruitment Status.

AI Recommendation

The supported values are:

  • Recommended
  • Consider
  • Not Recommended

These are AI outputs. They are not pipeline stages and cannot replace recruiter judgment.

Recruitment Status

Recruiters can update a candidate's status from the ranking view. Multiple selected candidates can also receive one bulk status update. Bulk updates are limited to a safe batch size.

Status changes are recorded in candidate history and appear in activity and historical analytics.

Compare and Export

  • Select suitable candidates and click Compare Candidates for a side-by-side AI comparison.
  • From the generated comparison, click Download PDF or Download Excel.
  • Click Export Excel to download candidate results for authorized internal use.
  • Click View Pipeline to manage candidates by Recruitment Status.

Experience Feedback Before Pipeline

After at least one candidate has been processed, each user is asked to submit brief experience feedback before opening the Pipeline for that job for the first time.

  • The feedback is required only once for each user and job.
  • Other workspace members submit their own feedback once.
  • The three ratings cover overall experience, ease of use, and processing speed.
  • The Tell us more description is mandatory; feature and issue fields remain optional.
  • Feedback does not change candidate scores, AI Recommendation, ranking, quota, or Recruitment Status.
  • Selecting Not now keeps you on Candidate Ranking; Pipeline becomes available after the feedback is submitted.

Free Trial CV Limit

A Free Trial workspace can upload up to 10 CVs to each job. The limit is reserved before expensive AI processing begins. A failed operation releases its reservation where applicable.

Responsible Use

  • Verify extracted information against the original resume.
  • Review explanations, not only the overall score.
  • Do not reject a candidate solely because of AI output.
  • Follow applicable privacy, retention, and anti-discrimination policies.

Continue to Candidate Directory for workspace-wide search or Candidate Profile for deeper review.

← Previous

Job Management

Next →

Candidate Directory