Smith B.

Key Skills

Software

Other

Top Subject Matter

LLM Evaluation and Annotation

NLP Model Training and Evaluation

Code Review and Security Labeling

Top Data Types

Image

Text

Computer Code Programming

Top Task Types

Computer Programming/Coding

Red Teaming

Text Summarization

Question Answering

Text Generation

Freelancer Overview

LLM Training Data Annotation and Response Evaluation. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Other. Education includes Bachelor of Engineering, University of International Relations (2026). AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Evaluation, Rating, and Classification.

Labeling Experience

LLM Training Data Annotation and Response Evaluation

Text

As an LLM data annotator and response evaluator, I created templates for instruction-following, code generation, debugging, tool-use reasoning, and safety review. I labeled model outputs for factuality, relevance, completeness, reasoning clarity, code correctness, hallucination risk, safety, and formatting compliance. I applied preference labeling criteria, wrote ranking rationales, and identified incomplete or unsafe responses. • Built reusable annotation rubrics with pass/fail standards, rejection reasons, and reviewer calibration examples. • Audited AI agent tool-use traces and labeled tool selection, unsupported claims, and misinterpretations. • Provided structured feedback and scoring for ambiguous or low-quality outputs. • Maintained label consistency and thorough documentation throughout the workflow.

2025 - Present

Data Processing and Annotation Workflow Support

OtherTextData Collection

In this workflow support and annotation role, I used Python, Pandas, and NumPy for dataset inspection, cleaning, label normalization, and quality statistics. I maintained reproducible annotation scripts and documented evaluation workflows using Git, Linux, and Docker. I organized model evaluation and labeling records to enable consistent review and revision. • Supported reproducibility and quality control in annotation processes. • Developed infrastructure for structured error tracking and rationales. • Ensured high-quality label normalization and dataset reliability. • Utilized internal scripts and automation tools to streamline annotation pipelines.

2024 - Present

NLP Dataset Labeling and Model Evaluation

OtherTextClassification

In this NLP dataset labeling and evaluation role, I worked on text classification and named entity recognition experiments using Kaggle and Google Colab. I cleaned and normalized text data, handled label splits, and reviewed ambiguous samples during model error analysis. I created instruction-style samples for LLM fine-tuning and evaluated QA outputs for consistency and hallucination risk. • Built domain-specific prompt-response pairs and negative examples for training. • Evaluated retrieval-augmented QA outputs using rubrics for factual quality. • Translated model errors into new labeling and data-improvement actions. • Improved labeling guidelines and annotation practices based on experiment findings.

2025 - 2025

Code, Security, and Red-Team Data Labeling

Other

In this code and security-focused labeling task, I reviewed AI-generated code for correctness, behavior, dependencies, tests, and security vulnerabilities. I designed red-team cases for prompt injection, unsafe scripts, data leakage, and insecure recommendations. I combined cybersecurity principles with AI data annotation to ensure safe and robust labeling. • Labeled high-risk outputs and wrote detailed safety comments for reviewers. • Participated in incident support and abnormal data detection during major events. • Used a security-first approach for labeling edge cases and ambiguous outputs. • Contributed to identifying errors in dependency and code generation cases.

2024 - 2025

Education

Work History