Shubham Wankhade

Key Skills

Software

Appen

Clickworker

Data Annotation Tech

Humanatic

LabelImg

Labelbox

OneForma

Micro1

Mindrift

Remotasks

Toloka

Telus

Surge AI

Mercor

Top Subject Matter

AI-generated code evaluation

LLM output ranking

Text and NLP data for LLM model training

Top Data Types

Text

Computer Code Programming

Image

Top Task Types

RLHF

Computer Programming/Coding

Prompt + Response Writing (SFT)

Diagnosis

Text Generation

Text Summarization

Fine-tuning

Question Answering

Evaluation/Rating

Data Collection

Function Calling

Transcription

Point/Key Point

Freelancer Overview

AI Code Evaluator (Code Preferences). Core strengths include Mercor and Telus. AI-training focus includes data types such as Computer Code, Programming, and Text and labeling workflows including Evaluation, Rating, and Data Collection.

Labeling Experience

Prompt response review

Scale AITextRLHFComputer Programming Coding

The Prompt Response Review project focused on evaluating and labeling large-scale text datasets generated from AI model responses in the domain of computer programming. The primary objective was to enhance model alignment, instruction-following capabilities, and response quality through Reinforcement Learning from Human Feedback (RLHF) and Supervised Fine-Tuning (SFT) tasks. The project involved critical assessment of AI responses to technical prompts related to programming, debugging, code review, and logical reasoning.

The Prompt Response Review project focused on evaluating and labeling large-scale text datasets generated from AI model responses in the domain of computer programming. The primary objective was to enhance model alignment, instruction-following capabilities, and response quality through Reinforcement Learning from Human Feedback (RLHF) and Supervised Fine-Tuning (SFT) tasks. The project involved critical assessment of AI responses to technical prompts related to programming, debugging, code review, and logical reasoning.

2024 - 2025

Backend Developer

Scale AIComputer Code ProgrammingDiagnosis

Work experience in Travel & Hospitality domain and worked for global projects like Amadeus and TravelClick. • Work experience in metaverse project model for future cities NEOM. • Worked on Core project module ,Vital module for Travel & Hospitality functions ARI(i.e.Availability,Rate,Inventory) processing, Booking, Services & Promotions etc. • Developed REST Web Services with 360 degree prespective from documentation to final deployment • Handling client communication to understand project requirement and devise solutions accordingly. • Migrate applications from ANT to MAVEN • Migrate GDS and Meta applications to rest-to-ih-adapter to retire CA-Gateway.

Work experience in Travel & Hospitality domain and worked for global projects like Amadeus and TravelClick. • Work experience in metaverse project model for future cities NEOM. • Worked on Core project module ,Vital module for Travel & Hospitality functions ARI(i.e.Availability,Rate,Inventory) processing, Booking, Services & Promotions etc. • Developed REST Web Services with 360 degree prespective from documentation to final deployment • Handling client communication to understand project requirement and devise solutions accordingly. • Migrate applications from ANT to MAVEN • Migrate GDS and Meta applications to rest-to-ih-adapter to retire CA-Gateway.

2023 - 2025

AI Code Evaluator (Code Preferences)

Mercor

As an AI Code Evaluator at Mercor, I evaluated AI-generated code outputs for correctness, logical reasoning, and efficiency. My responsibilities included ranking multiple LLM responses based on quality and adherence to instructions. I also provided structured feedback to improve model accuracy and addressed logical inconsistencies. • Evaluated code in Python, Java, and C++. • Identified edge cases and hallucinations in model outputs. • Compared and rated LLM-generated code for quality. • Ensured adherence to provided guidelines for each evaluation.

As an AI Code Evaluator at Mercor, I evaluated AI-generated code outputs for correctness, logical reasoning, and efficiency. My responsibilities included ranking multiple LLM responses based on quality and adherence to instructions. I also provided structured feedback to improve model accuracy and addressed logical inconsistencies. • Evaluated code in Python, Java, and C++. • Identified edge cases and hallucinations in model outputs. • Compared and rated LLM-generated code for quality. • Ensured adherence to provided guidelines for each evaluation.

Not specified

Education

Work History