Debjani C.

Key Skills

Software

Other

Telus

Top Subject Matter

Enterprise LLM evaluation (RLHF & safety)

Multimodal dataset annotation and multilingual QA

AI prompt/query annotation and QA

Top Data Types

Text

Image

Top Task Types

RLHF

Question Answering

Prompt + Response Writing (SFT)

Freelancer Overview

AI/LLM Evaluator (Outlier). Brings 14+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other and Telus. Education includes Master of Arts, Indira Gandhi National Open University (IGNOU) (2012) and Bachelor's Degree, S.T.C.M. College, University of Calcutta (2003). AI-training focus includes data types such as Text and Image and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

AI Annotator / Prompter (Mindrift)

OtherTextPrompt + Response Writing (SFT)

Reviewed AI outputs and created structured prompts for domain-specific tasks. Performed prompt evaluation to improve clarity and accuracy of generated results for training or task execution. Supported AI tutoring-style quality improvement by refining prompts and checking output quality.• Conducted prompt evaluation and output review for accuracy• Wrote structured prompts aligned to domain-specific requirements• Improved clarity and correctness of AI-generated responses• Performed iterative refinements to meet task quality expectations

2025 - Present

AI Prompter (TELUS Digital)

TelusTextQuestion Answering

Labeled datasets and performed quality assurance on AI prompts and query responses for downstream model improvement. Conducted query vetting and applied consistency checks against annotation guidelines across tasks. Supported evaluation of prompt quality and response appropriateness through structured labeling workflows.• Performed query vetting to filter and assess user inputs• Conducted SxS comparison and multi-turn evaluation activities• Performed instruction-following QA against provided guidelines• Ensured consistency across annotations and quality checks

2025 - Present

AI Annotator (Innodata / AI Superintelligence)

OtherImage

Annotated and evaluated multimodal image/video datasets to maintain high QA standards and accuracy. Applied structured evaluation frameworks, including ranking and instruction-following validation, to assess response quality across multilingual datasets. Maintained approximately a 2.8/3 QA score while handling high-volume annotation workflows.• Performed ranking-based evaluation of model responses• Validated instruction-following and conducted quality checks• Annotated image/video content for multimodal dataset QA• Managed high-volume workflows (5–6 hrs/day) to meet throughput goals

2025 - Present

AI/LLM Evaluator (Outlier)

OtherTextRLHF

Evaluated real-user queries and model outputs to improve enterprise AI reliability and safety. Performed query vetting, SxS comparisons, and multi-turn evaluation to identify hallucinations, inconsistencies, and bias. Made independent PASS/FAIL decisions based on quality, instruction adherence, and structured reasoning.• Conducted SxS comparisons across model responses to assess correctness• Reviewed multi-turn conversations for instruction following and reasoning quality• Applied hallucination detection and bias checks for safety evaluation• Issued PASS/FAIL judgments to support QA and model improvement

2026 - 2026

Education

Work History