Shanmukh Kiran Sagar

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Top Label Types

RLHF

Evaluation Rating

Function Calling

Prompt Response Writing SFT

Freelancer Overview

I am a software engineering student with hands-on experience in AI evaluation, data annotation, and training data workflows. My background includes RLHF-based code review, where I evaluated agent-generated outputs for correctness, reasoning quality, and task alignment, as well as defining gold-standard behaviors and analyzing failure modes. I am skilled in Python, familiar with Java and Rust, and comfortable working with Git, Docker, and structured data formats like JSON and YAML. My experience also includes ML annotation tasks, prompt refinement, and bias testing, giving me a strong foundation in quality-driven AI evaluation. I am passionate about contributing to robust AI systems and am open to part-time or project-based opportunities in data labeling and AI training data.

Labeling Experience

LLM Alignment & RLHF-Based Code Evaluation Project

LabelboxComputer Code ProgrammingRLHFEvaluation Rating

Worked on AI alignment and RLHF-based evaluation workflows for code-generation agents.Reviewing and comparing agent-generated code outputs for correctness, reasoning depth, and instruction alignment, Defining gold-standard behavioral expectations for multi-step execution workflows, Performing structured quality scoring and human preference ranking, Analyzing failure modes, execution logs, and reasoning traces, Iterating on prompts and evaluation instructions to improve robustness and reduce ambiguity Project Scope: Evaluated complex code-generation tasks across varied technical domains Applied structured criteria using JSON-based evaluation formats Conducted consistency checks to maintain scoring reliability Quality Measures: Multi-pass review system Cross-checking reasoning chains against task constraints Bias and logical coherence validation Focused on improving model alignment, reasoning reliability, and instruction-following behavior.

Worked on AI alignment and RLHF-based evaluation workflows for code-generation agents.Reviewing and comparing agent-generated code outputs for correctness, reasoning depth, and instruction alignment, Defining gold-standard behavioral expectations for multi-step execution workflows, Performing structured quality scoring and human preference ranking, Analyzing failure modes, execution logs, and reasoning traces, Iterating on prompts and evaluation instructions to improve robustness and reduce ambiguity Project Scope: Evaluated complex code-generation tasks across varied technical domains Applied structured criteria using JSON-based evaluation formats Conducted consistency checks to maintain scoring reliability Quality Measures: Multi-pass review system Cross-checking reasoning chains against task constraints Bias and logical coherence validation Focused on improving model alignment, reasoning reliability, and instruction-following behavior.

2025

Education

Work History