Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
Inscribe seeks a seasoned ML researcher to own fraud detection research end-to-end, focusing on image tampering and AI-generated documents. You will experiment with LLMs and VLMs, evaluate on messy unlabeled data, and collaborate with engineers to deploy robust detectors in production environments.
Experience in computer vision, LLM/VLM evaluation, and Python/SQL within a real codebase is crucial. AWS Bedrock and Anthropic platforms are part of the tooling landscape.
The role asks for someone to own fraud detection research end to end, with a strong focus on image files such as tampered images and AI generated documents. A strong candidate brings several years of applied ML work with models running in production, solid statistics for evaluating results on messy unlabelled data, and hands on computer vision plus LLM and VLM experience. Comfort with Python and SQL inside a real codebase, working alongside engineers, matters more than a long list of tools.
AI is both the tool and the target here, since the work involves building detectors for AI generated documents while also using LLMs and VLMs to catch fraud. Much of the routine model comparison and pipeline building can be sped up, but judging when a model is the wrong choice and reading real documents for failure modes stays human. The role sits where new AI methods meet careful evaluation against real customer traffic.
A person in this role would gain direct practice evaluating and comparing LLMs and VLMs, running prompt and agent experiments, and building pipelines on AWS Bedrock and Anthropic platforms. They would also learn image forensics and tamper detection, which builds a rare skill for spotting synthetic or edited documents. Working with unlabelled production data and low fraud rates adds experience in rigorous measurement that transfers well to other risk work.
This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.
AI-powered document fraud detection and risk intelligence for banks, credit unions and fintech lenders.