Technical Data Annotator

Location: Remote Project duration: Approximately 12 weeks Compensation: $6.50 USD per hour Employment type: Independent contractor / project-based Hours: To be confirmed About the Role We are seeking technically literate data annotators to support an AI safety project focused on improving the guardrail systems of an enterprise coding assistant. In this role, you will review and annotate multi-turn conversations between users and an AI coding assistant. These conversations may contain natural language, source code, command-line instructions, configuration files, system logs, and other technical content. Your primary responsibility will be to identify potential safety risks related to content moderation and prompt injection while considering the context and intent of each interaction. We will provide project-specific guidelines and training.

Responsibilities

Key Responsibilities Review multi-turn, code-related conversations involving user requests and AI-generated responses. Interpret code snippets, code changes, system logs, command-line instructions, and structured tool outputs. Identify and label potential prompt-injection attempts, such as instructions intended to override system rules or hidden malicious content. Identify content-moderation risks, including hateful, harmful, or illicit content. Evaluate the context and intent of technical language rather than relying only on individual keywords. For example, terms such as “kill a process� or “execute a payload� may be legitimate in a programming context. Replace masked or redacted placeholders with realistic, neutral synthetic information when required. Follow detailed annotation guidelines and apply multi-level taxonomies consistently. Participate in regular calibration sessions and meet established quality and accuracy standards. Escalate unclear or ambiguous cases according to project procedures.

Qualifications

Required Qualifications Bachelor’s degree in Computer Science, Software Engineering, Information Technology or another related STEM field, or equivalent technical training and demonstrable experience. Familiarity with at least one programming language, such as Python, Java, JavaScript, C/C++, Go or SQL. Ability to read and interpret basic code, code changes, configuration files, stack traces, command-line instructions, and structured technical outputs. Basic understanding of application-security concepts, such as injection attacks, authentication bypass, and backdoors. Professional-level English reading and comprehension skills. Strong analytical and contextual reasoning abilities. Excellent attention to detail and ability to apply complex guidelines consistently. Availability for the expected duration and weekly hours of the project. Preferred Qualifications Previous experience in data annotation, content moderation, AI evaluation, or RLHF projects. Familiarity with responsible AI, guardrail systems, or prompt-injection detection. Experience evaluating chatbots, multi-turn conversations, or developer-facing AI tools. Experience using annotation platforms such as Label Studio, Prodigy or similar tools. Understanding of personally identifiable information, data anonymization and responsible data handling.

Conditions

Important Information Selected candidates will be required to complete a technical screening assessment. This project may involve reviewing potentially harmful, offensive, or sensitive text as part of content-moderation tasks. You must participate in project training and calibration sessions.