Job Description: AI Data Review & Training Intern
Role: AI Data Review & Training Intern Department: AI Automation / Operations Location: Sec-62, Golf course extn, Gurgaon Internship Duration: 6 months Working Hours: 8
Role Overview
We are looking for a detail-oriented intern to support the development of an AI-based email automation and data-extraction system. The intern will review information extracted by Large Language Models (LLMs), correct inaccurate outputs, and create high-quality training examples to improve model performance.
This role combines data review, business-process understanding, structured data validation, and basic AI operations. Prior machine-learning development experience is not mandatory, but the candidate should be comfortable working with structured data and learning AI-assisted workflows.
Key Responsibilities
· Review data extracted by LLMs from emails, email threads, tables, documents, and attachments.
· Compare model predictions with the original source and identify missing, incorrect, or misclassified information.
· Correct extracted fields such as:
. Customer name and customer code
. Invoice and document numbers
. Payment amount and payment date
. UTR and cheque numbers
. Deduction amount and reason
. Commitment or PTP details
. Email category, disposition, and required action
· Create accurate, consistent, and complete training examples using the provided review application.
· Validate whether the model selected the correct email section or table row.
· Identify duplicate, incomplete, ambiguous, or unsuitable training examples.
· Follow annotation and review guidelines to maintain consistent data quality.
· Record recurring model errors and extraction patterns.
· Escalate unclear business scenarios to the project lead instead of making unsupported assumptions.
· Support quality checks before reviewed data is approved for model training.
· Maintain the confidentiality of customer, financial, and email data.
· Help document new scenarios, field definitions, and review rules discovered during the process.
Required Skills
· Strong attention to detail and accuracy.
· Good written English and reading comprehension.
· Comfortable reviewing long email threads and structured financial data.
· Working knowledge of Microsoft Excel or Google Sheets.
· Ability to understand tables, dates, amounts, customer references, and transaction details.
· Ability to follow documented rules and apply them consistently.
· Basic understanding of AI, LLMs, or data annotation.
· Good analytical and problem-solving skills.
· Ability to work independently and escalate uncertainties appropriately.
Preferred Qualifications
· Currently pursuing or recently completed a degree in:
. Computer Science
. Information Technology
. Data Science
. Business Analytics
. Finance or Commerce
. Operations or a related discipline
· Exposure to data annotation, quality assurance, OCR, NLP, or AI tools.
· Familiarity with financial operations, collections, accounts receivable, or customer emails.
· Basic knowledge of Python, SQL, JSON, or databases is helpful but not mandatory.
What Success Looks Like
· Reviewed examples accurately match the source emails and attachments.
· Important fields are not skipped or populated with unsupported values.
· Corrections follow the prescribed format and annotation guidelines.
· Ambiguous examples are flagged rather than guessed.
· Recurring model errors are documented clearly.
· Training data remains consistent across reviewers and batches.
· Assigned review batches are completed within agreed timelines without compromising quality.
Key Performance Indicators
· Review accuracy and quality-audit score.
· Number of valid training examples completed.
· Percentage of examples requiring rework.
· Consistency in applying annotation guidelines.
· Ability to identify missing or incorrect model predictions.
· Timeliness of review-batch completion.
· Quality of documented observations and escalations.
Pay: ₹5,000.00 - ₹10,000.00 per month
Work Location: In person
