For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
A

Abhinandan Pathak

LLM Response Evaluation & RLHF Preference Labeling

Canada flagBrampton, Canada
$35.00/hrExpert

Key Skills

Software

No software listed

Top Subject Matter

Large Language Model (LLM) code generation and evaluation
Compiler error trace annotation and error classification
Source code security vulnerability classification (CVE)

Top Data Types

TextText
DocumentDocument
ImageImage

Top Task Types

Text Generation
Question Answering
Text Summarization
Fine Tuning
Data Collection
Prompt Response Writing SFT
Computer Programming Coding
Classification
Segmentation
RLHF
Entity Ner Classification

Freelancer Overview

LLM Response Evaluation & RLHF Preference Labeling. Brings 5+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Internal and Proprietary Tooling. Education includes Postgraduate Degree, Algoma University (2023) and Bachelor of Business Administration, Galgotias University (2021). AI-training focus includes data types such as Computer Code and Programming and labeling workflows including RLHF, Entity (NER) Classification, and Classification.

ExpertEnglish

Labeling Experience

Frontend Code Accessibility & UX Labeling

Classification
Annotated HTML, CSS, and React code snippets for accessibility and WCAG compliance. Provided labels on accessibility violations for 850+ samples to support frontend improvement tools. Labeled data was used to train AI systems for automatic accessibility feedback in frontend code. • Labeled code for accessibility violations. • Supported WCAG compliance detection. • Processed diverse frontend codebases. • Improved AI tools for frontend UX accessibility.

Annotated HTML, CSS, and React code snippets for accessibility and WCAG compliance. Provided labels on accessibility violations for 850+ samples to support frontend improvement tools. Labeled data was used to train AI systems for automatic accessibility feedback in frontend code. • Labeled code for accessibility violations. • Supported WCAG compliance detection. • Processed diverse frontend codebases. • Improved AI tools for frontend UX accessibility.

2025 - 2026

SQL Query Validation & Error Labeling

Classification
Validated and labeled SQL queries for AI-driven query optimization tools, identifying errors and performance issues. Processed over 900 queries, providing precise feedback on query structure, correctness, and optimization needs. Supported development of database query improvement AI systems. • Reviewed SQL queries for errors. • Labeled performance and logic issues. • Supplied annotated samples for AI optimization models. • Improved accuracy of AI-assisted database tools.

Validated and labeled SQL queries for AI-driven query optimization tools, identifying errors and performance issues. Processed over 900 queries, providing precise feedback on query structure, correctness, and optimization needs. Supported development of database query improvement AI systems. • Reviewed SQL queries for errors. • Labeled performance and logic issues. • Supplied annotated samples for AI optimization models. • Improved accuracy of AI-assisted database tools.

2025 - 2025

Code Review Ranking & Preference Labeling

RLHF
Evaluated and ranked AI-generated code completions across Python, Java, and C++ for RLHF training. Focused on correctness, efficiency, and readability, contributing over 1,200 ranked pairs for code generation improvement. Labeled outputs supported the development of more aligned code models. • Labeled code completions with preference scores. • Improved model quality for LLM code tasks. • Contributed to RLHF code training datasets. • Provided ranking feedback for pipeline optimization.

Evaluated and ranked AI-generated code completions across Python, Java, and C++ for RLHF training. Focused on correctness, efficiency, and readability, contributing over 1,200 ranked pairs for code generation improvement. Labeled outputs supported the development of more aligned code models. • Labeled code completions with preference scores. • Improved model quality for LLM code tasks. • Contributed to RLHF code training datasets. • Provided ranking feedback for pipeline optimization.

2025 - 2025

Python Code Quality & Bug Detection Labeling

Classification
Reviewed and labeled Python code snippets to train bug detection AI models. Identified logic flaws, syntax errors, and security issues, maintaining high labeling quality and accuracy. Labeled over 1,000 samples ensuring effective data for model training. • Supported bug detection model training. • Maintained 96% labeling accuracy rates. • Tagged samples for syntax and logic errors. • Enhanced quality of labeled bug detection data.

Reviewed and labeled Python code snippets to train bug detection AI models. Identified logic flaws, syntax errors, and security issues, maintaining high labeling quality and accuracy. Labeled over 1,000 samples ensuring effective data for model training. • Supported bug detection model training. • Maintained 96% labeling accuracy rates. • Tagged samples for syntax and logic errors. • Enhanced quality of labeled bug detection data.

2025 - 2025

Code Readability & Maintainability Scoring for LLM Fine-Tuning

Scored and ranked code samples from Python, TypeScript, and Java for readability and maintainability using a 10-point rubric. Evaluated code for DRY, decomposition, and cyclomatic complexity requirements to fine-tune LLMs. The effort improved developer productivity in A/B testing with real-world engineers. • Ranked 2,200+ code samples for maintainability. • Used a detailed scoring system for fairness. • Supported LLM fine-tuning for readable code outputs. • Boosted developer scores by 23% in testing.

Scored and ranked code samples from Python, TypeScript, and Java for readability and maintainability using a 10-point rubric. Evaluated code for DRY, decomposition, and cyclomatic complexity requirements to fine-tune LLMs. The effort improved developer productivity in A/B testing with real-world engineers. • Ranked 2,200+ code samples for maintainability. • Used a detailed scoring system for fairness. • Supported LLM fine-tuning for readable code outputs. • Boosted developer scores by 23% in testing.

2024 - 2025

Education

G

Galgotias University

Bachelor of Business Administration, Business Administration (Marketing)

Bachelor of Business Administration
2019 - 2021
A

Algoma University

Postgraduate Degree, Human Resource and Business Management

Postgraduate Degree
2023

Work History

6

6IXTYWINGS

Customer Relations & Operations Manager

Brampton
2023 - Present
6

6IXTYWINGS

Operations & Service Representative

Brampton
2023 - 2023