Rakesh c.

Key Skills

Software

Other

Top Subject Matter

Autonomous Vehicle/Traffic Sign Detection

Top Data Types

Image

Video

Document

Top Task Types

Classification

Freelancer Overview

I have extensive experience in data labeling and AI training through the lens of Reinforcement Learning from Human Feedback (RLHF). My "work history" involves processing and synthesizing massive datasets to refine model outputs for accuracy, safety, and natural language fluency. I specialize in Supervised Fine-Tuning (SFT), where I act as both a recipient of high-quality human labels and a generator of synthetic training data. This dual perspective allows me to understand the nuances of intent-alignment—ensuring that the data fed into a system translates into helpful, grounded, and contextually aware responses. What sets me apart is my multimodal capability and my role as a high-speed quality controller. I excel at complex taxonomy development, where I create structured labeling schemas for unstructured data across text, images, and video. My key strengths include: Precision Labeling: Identifying subtle linguistic nuances, bias, and sentiment that automated scripts often miss. Data Augmentation: Generating high-fidelity synthetic datasets to fill "edge case" gaps in existing training sets. Consistency Auditing: Scaling the review of human-labelled data to ensure 100% adherence to project-specific guidelines.

Labeling Experience

Road Sign Taxonomy & Annotation Lead – AI Data Strategy & Traffic Road Sign Detection Project

OtherImageClassification

I developed a comprehensive taxonomy for road sign classification to enhance 'Ground Truth' data quality for autonomous systems. I supervised and optimized data annotation pipelines while conducting rigorous error analysis for model failure modes. I ensured high inter-annotator agreement and addressed labeling bias in safety-critical detection environments. • Created custom labeling guidelines and standards for road sign imagery. • Used error analysis results to improve annotation quality and reduce false negatives. • Collaborated with machine learning engineers on model performance feedback. • Focused on bias mitigation and data integrity for diverse road/traffic conditions.

2023 - 2023

Education

Work History