Open to opportunities

Mohini Yadav

@mohiniyadav

Message

I evaluate and improve AI models through RLHF, prompt review, relevance rating, and data annotation.

India

Message

What I'm looking for

I'm looking for AI training, LLM evaluation, data annotation, search relevance, or quality-review work where I can improve model accuracy, safety, and alignment through structured, guideline-driven feedback.

At Innodata, I evaluate LLM outputs for accuracy, relevance, safety, and factual correctness, applying RLHF feedback to strengthen model alignment. I also conduct video annotation, bounding-box labeling, prompt analysis, and quality control across large-scale AI training workflows.

Across Welocalize, Mercor, and MopshyAI, I've ranked AI responses, reviewed prompts, assessed audio outputs, moderated content, rated search relevance, and annotated user–Siri conversations. My work combines structured rationales, cultural nuance in Hindi, and guideline-driven review to improve model quality.

With a B.Sc. in Mathematics, I also bring experience delivering Physics solutions and creating accurate English educational content.

Experience

Work history, roles, and key accomplishments

Current

AI/LLM Analyst

Current

Innodata

May 2026 - Present (3 months)

Evaluate LLM outputs for accuracy, relevance, safety, and factual correctness, applying RLHF principles. Perform video annotation and bounding-box labeling, conduct prompt engineering, and lead quality control.

LLM Evaluation Video Annotation Bounding Boxes Prompt Engineering Quality Control HumanSignal Multimango Feather Access

Current

Quality Control Specialist

Current

Welocalize

Feb 2026 - Present (6 months)

Worked as Hindi Culture Expert, Data Trainer, and AI Content Reviewer, evaluating AI-generated content for cultural nuance, audio quality, and factual correctness. Performed search and data labeling, prompt review, and RLHF feedback.

Audio Evaluation Data Labeling Content Moderation Prompt Review Google Sheets

Generalist Expert

Mercor

May 2026 - Jul 2026 (2 months)

Engineered and reviewed prompts to elicit higher-quality model responses. Ranked AI-generated responses by quality and user intent, writing detailed rationales to refine accuracy and relevance.

Prompt Engineering Response Ranking

AI Data Annotator

MopshyAI

Mar 2026 - Jun 2026 (3 months)

Trained Siri using iQuota.com, an Apple annotation tool, working on transcription and semantic annotation of user–Siri conversations. Applied filler words and natural speech patterns to improve conversational accuracy.

Data Annotation Transcription IQuota