Riya Manro

Key Skills

Software

Labelbox

Label Studio

OneForma

Telus

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Evaluation/Rating

Fine-tuning

Prompt + Response Writing (SFT)

RLHF

Freelancer Overview

I have hands-on experience working as an AI trainer focused on LLM (Large Language Model) evaluation and prompt creation, with a specialization in both English and Hindi tasks. My work has involved comparing model outputs, crafting high-quality prompts, and annotating responses for accuracy, clarity, and relevance. In addition to text-based tasks, I have also annotated images, videos, and bounding boxes, contributing to computer vision datasets with precise labeling and attention to detail. Alongside AI training and data annotation, I have a background in UI development and technical writing. I've authored multiple technical articles for GeeksforGeeks and developed web interfaces using technologies like ReactJS and JavaScript. This blend of technical and communication skills allows me to understand AI systems deeply and contribute effectively to projects across NLP and computer vision domains.

Labeling Experience

Penguin - Annotation Project

Internal Proprietary ToolingImageBounding BoxQuestion Answering

As part of a large-scale multimodal AI training project, I annotated over 1,500 images by creating precise bounding boxes around key objects, labeling each with accurate tags, and evaluating the quality and completeness of annotations. I also generated over 2,000 creative, context-based questions based on image content to support visual reasoning in AI models (e.g., “What is the child reaching for on the table?”). My responsibilities included reviewing image clarity, ensuring labeling consistency, and maintaining high annotation standards, with over 95% bounding box accuracy (IoU) and a 90%+ acceptance rate for question quality.

As part of a large-scale multimodal AI training project, I annotated over 1,500 images by creating precise bounding boxes around key objects, labeling each with accurate tags, and evaluating the quality and completeness of annotations. I also generated over 2,000 creative, context-based questions based on image content to support visual reasoning in AI models (e.g., “What is the child reaching for on the table?”). My responsibilities included reviewing image clarity, ensuring labeling consistency, and maintaining high annotation standards, with over 95% bounding box accuracy (IoU) and a 90%+ acceptance rate for question quality.

2025 - 2025

RLHF

Label StudioTextFine TuningPrompt Response Writing SFT

I worked on an RLHF (Reinforcement Learning with Human Feedback) project focused on improving the alignment and quality of large language model outputs. My role involved evaluating and ranking over 3,000 sets of model responses based on relevance, clarity, accuracy, tone, and safety. I provided detailed written rationales for each ranking to support model fine-tuning and training. The tasks spanned diverse domains including general knowledge, creative writing, technical queries, and safety-sensitive topics. I maintained a reviewer agreement rate of over 95%, consistently meeting quality and consistency standards in a fast-paced, high-volume evaluation environment.

I worked on an RLHF (Reinforcement Learning with Human Feedback) project focused on improving the alignment and quality of large language model outputs. My role involved evaluating and ranking over 3,000 sets of model responses based on relevance, clarity, accuracy, tone, and safety. I provided detailed written rationales for each ranking to support model fine-tuning and training. The tasks spanned diverse domains including general knowledge, creative writing, technical queries, and safety-sensitive topics. I maintained a reviewer agreement rate of over 95%, consistently meeting quality and consistency standards in a fast-paced, high-volume evaluation environment.

2024 - 2024

Education

Work History