Fraud/ML Engineer

You will make sure that only authentic, high-signal, human data gets through a system that processes millions of files every day. You'll build AI-generated image and video detection, reverse image search and plagiarism rejection tools, duplicate fingerprinting using vector and perceptual hashing, copyright risk detection, EXIF and metadata tampering detection, fraud network and device clustering systems, and human-in-the-loop verification pipelines. You'll work on adversarial machine learning problems at scale rather than academic benchmarks, catching increasingly complex fraud attempts across a massive volume of daily uploads.

Responsibilities

• Build AI-generated image and video detection systems
• Build reverse image search and internet plagiarism rejection tools
• Implement duplicate fingerprinting using vector and perceptual hashing
• Build copyright risk detection systems
• Detect EXIF and metadata tampering
• Build fraud network and device clustering systems
• Build human-in-the-loop verification pipelines

Requirements

• 3+ years in computer vision / ML (PyTorch or TensorFlow)
• Production ML deployment experience
• Strong SQL / PostgreSQL skills
• Experience with vector search (FAISS, pgvector, Pinecone)
• Image processing (OpenCV, PIL)
• Comfort shipping backend systems (TypeScript/Deno or similar)
• Deepfake detection experience is a bonus
• Reverse image search systems experience is a bonus
• Copyright detection pipelines experience is a bonus
• Trust & Safety infrastructure experience is a bonus

Benefits

• Equity compensation