Shubham Shukla

Key Skills

Software

Mercor

OneForma

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Audio Recording

Evaluation Rating

Object Detection

Prompt Response Writing SFT

Text Generation

Freelancer Overview

I am an AI Data Labeling and AI Training specialist with hands-on experience in text, image, and code-based annotation tasks. I have contributed to multiple AI training workflows including LLM evaluation, prompt-response quality review, text classification, and image annotation for computer vision models. My background in computer science and strong programming skills allow me to deliver high-accuracy annotations while strictly following project guidelines and quality benchmarks. I have worked on tasks such as prompt and response writing (SFT), evaluation and rating of AI-generated outputs, entity recognition, object detection, and segmentation. Additionally, I bring experience in Python-based workflows, data validation, and QA processes, enabling me to understand model behavior and improve training data quality. I am detail-oriented, consistent, and comfortable working on large-scale AI training projects with minimal supervision.

Labeling Experience

LLM Evaluation & Prompt-Response Annotation

Scale AITextText GenerationRLHF

Worked on LLM training workflows involving evaluation and ranking of AI-generated responses based on accuracy, relevance, tone, and safety. Performed prompt-response writing (SFT), question-answer validation, and comparative evaluation to improve model alignment. Followed detailed annotation guidelines, applied consistency checks, and maintained high quality standards across large text datasets.

Worked on LLM training workflows involving evaluation and ranking of AI-generated responses based on accuracy, relevance, tone, and safety. Performed prompt-response writing (SFT), question-answer validation, and comparative evaluation to improve model alignment. Followed detailed annotation guidelines, applied consistency checks, and maintained high quality standards across large text datasets.

2024 - 2025

Education

Work History