About the Role
As an ML Engineer specializing in Fraud Detection & Data Quality, you will be responsible for ensuring that only authentic, high-signal, human data enters Kled's system. You will build and implement advanced adversarial ML systems at scale to process millions of daily uploads.
Responsibilities
- Develop AI-generated image & video detection systems.
- Implement reverse image search & internet plagiarism rejection mechanisms.
- Create duplicate fingerprinting solutions using vector and perceptual hashing.
- Build copyright risk detection tools.
- Design EXIF / metadata tampering detection systems.
- Develop fraud network & device clustering systems.
- Establish human-in-the-loop verification pipelines.
Requirements
- 3+ years of experience in computer vision / ML (PyTorch or TensorFlow).
- Proven production ML deployment experience.
- Strong proficiency in SQL / PostgreSQL.
- Experience with vector search technologies (FAISS, pgvector, Pinecone).
- Proficiency in image processing (OpenCV, PIL).
- Comfort with shipping backend systems (TypeScript/Deno or similar).
Bonus Skills
- Experience with deepfake detection.
- Knowledge of reverse image search systems.
- Familiarity with copyright detection pipelines.
- Experience with Trust & Safety infrastructure.
Current Stack
Backend
- PostgreSQL (Supabase) – managing hundreds of millions of media files.
- S3 storage.
- Deno / TypeScript edge functions.
- Python detection pipelines.
Frontend
- SwiftUI (migrating to Flutter).
- Internal verification tooling.
Growth Opportunity
You will join a team at the frontier of applied AI data infrastructure, with the chance to:
- Own core systems for one of the world's largest human data networks.
- Design infrastructure influencing next-gen AI model training data.
- Build at real scale with millions of daily uploads in adversarial environments.
- Collaborate with a team experienced in building marketplaces, AI systems, and products used by millions.