Staff Software Engineer, Trust & Safety

Design and implement detection systems to prevent abuse on an AI-native platform.

• •Replit's Trust & Safety team builds detection systems and automated responses to prevent abuse on an AI-native platform, focusing on LLM-specific attacks and large-scale operational defenses.
• •Key Responsibilities Design and implement LLM guardrails and AI-powered detection systems Build and operate abuse detection for phishing, cryptomining, account takeover, and fraud Design automated enforcement and own the full abuse response lifecycle Analyze attack patterns using BigQuery/Hex and extend internal detection tools Requirements 8+ years in security engineering, anti-abuse, or fraud detection Strong programming skills in Python and/or TypeScript Experience with SQL and data analysis at scale (BigQuery/Snowflake) Experience building or fine-tuning ML/LLM classifiers and familiarity with prompt injection