Video Transcript Editor - Accuracy, Speaker Labels, & Context

I’m looking for a highly detail oriented person to help edit and perfect AI generated transcripts of video courses. The videos have already been transcribed using AI and are approximately 95% accurate, so you will NOT be transcribing everything from scratch. Your job is to watch parts of each original video while reviewing the existing transcript and correct the remaining errors and missing context. This requires more than basic proofreading. The videos include a main speaker teaching an audience, interacting with audience members, answering questions, giving demonstrations, and sometimes using body language, tone, pauses, or visual actions that are important to understanding what he means. Your job will include: -Correct any words or sentences the AI transcribed incorrectly. -Correctly distinguish who is speaking and fix incorrect speaker labels. -Preserve the speaker’s actual meaning and wording rather than rewriting the material. -Add short notes when important visual/nonverbal context is necessary to understand what is happening (for example, pointing at someone, demonstrating a posture, deliberately changing his tone, or acting something out). -Add simple context labels when the type of content changes, such as direct teaching, audience interaction, demonstration, guided exercise, etc. -Remove obvious AI transcription errors or artifacts. -Flag anything genuinely unclear rather than guessing. The goal is extremely accurate, well structured transcripts that preserve not only what was said, but any important context needed to correctly understand the material. I already have several transcripts that I personally edited to the standard I want. These will be provided as examples, so you will have a clear reference for how the finished transcripts should look. I will provide the original videos and existing transcripts through Google Drive. What I’m looking for: -Excellent English comprehension -Very strong attention to detail -Ability to distinguish between multiple speakers -Ability to notice meaningful visual/nonverbal context -Good judgment about what information matters -Willingness to follow an established format consistently -Reliability and accuracy over rushing There are approximately 40-80 videos total, with several already completed. If we work well together, I would like the same person to complete the remaining collection. Please briefly tell me about any experience you have editing or verifying transcripts against original video/audio. Shortlisted applicants may be given a short sample so I can make sure we are a good fit before beginning the full project.