Fraud Engineer - AI Specialist

Obsidian; San Francisco (CA); On-site; USD 183,220 - 275,520; Part time

Job summary

Obsidian is collaborating with a leading AI research lab for a project involving Frontier Code Agents. Contributors will focus on evaluating and improving AI coding models relevant to fraud and risk engineering.

This role requires 2+ years in relevant fields and the ability to assess and utilize AI coding tools. Compensation is $400 per accepted task, typically taking 2–3 hours to complete.

Qualifications

• 2+ years of professional experience in fraud engineering or related fields.
• Ability to evaluate AI model-generated solutions in the context of fraud and risk.
• Experience with building and assessing fraud detection models.

Responsibilities

• Evaluate and improve frontier AI coding models via technical assessments.
• Complete complex fraud and risk engineering tasks using AI coding agents.
• Identify bugs and assess the strengths of different models.

Skills

Fraud engineering

Risk engineering

AI coding agents

Fraud detection systems

Tools

Cursor

Claude Code

Codex

Job description

About the Role

• Mercor is partnering with a leading AI research lab to support a Frontier Code Agents project.
• Contributors help evaluate and improve frontier AI coding models through structured technical assessments.
• The work focuses on realistic fraud, risk, and trust & safety engineering workflows and model evaluation.
• Spots are limited and filling quickly on a first come, first serve basis.

What You'll Do

• Use frontier AI coding agents to complete and evaluate complex fraud and risk engineering tasks.
• Review model-generated implementations involving fraud detection systems, risk scoring models, abuse prevention tooling, and transaction monitoring systems.
• Identify bugs, edge cases, and failure modes.
• Compare outputs from multiple frontier models and assess their strengths and weaknesses.
• Apply professional engineering judgment to realistic fraud and risk engineering scenarios.

Time Commitment

Sprint based project that runs in 12-24 hour stretches based on client requirement.

Compensation

• $400 per accepted task.
• Typical tasks take approximately 2–3 hours after ramp-up.
• Compensation is tied to accepted work.

Who Should Apply

• 2+ years of professional experience in fraud engineering, risk engineering, trust & safety, or related domains.
• Experience building fraud detection systems, risk scoring models, abuse prevention tooling, or transaction monitoring systems.
• Regular use of AI coding agents such as Cursor, Claude Code, Codex, Windsurf, Gemini CLI, or similar tools.
• Ability to evaluate model-generated solutions in fraud, risk, and trust & safety contexts.
• Experience in fintech, payments, marketplaces, or large-scale consumer platforms is preferred.