This initial endeavor is a stepping stone within a grander plan to create a robust deepfake detection framework.
Data Science: Evaluate the capabilities of 2-3 commercial deepfake detection APIs by sending requests to their provided endpoints with a subset of 500 items from our internal dataset. Evaluate, analyze and compare their detection results.
Research Engineering: Test 3-4 state-of-the-art deepfake detection algorithms (for example, as listed on https://paperswithcode.com/sota/deepfake-detection-on-fakeavceleb-1 ) and evaluate them.
You would need to test our internal datasets against these models, analyze and write reports, and eventually, deploy the model into production.
Having hands-on experience in working with research GitHub repositories and proficiently navigating through them is essential.
This job is already closed and no longer accepting applicants, sorry.