Audit Italian speech transcriptions, word-level timestamps, and alignment quality for AI training data. This remote contract role pays $39.50 per hour with a default commitment of 10 hours weekly.
About OpenTrain
OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI contracts contributors for specialized projects where careful human judgment helps improve modern artificial intelligence systems.
About AI Training Work
AI systems learn from examples prepared and reviewed by people. Audio contributors help models understand speech by transcribing recordings, checking spoken content, and evaluating whether time-aligned labels accurately match what was said.
This work is remote and can offer flexible, part-time opportunities. It is a practical way to build experience in a fast-growing technology field while contributing to how speech and language AI systems perform.
The Role
OpenTrain AI is seeking an Italian Speech Transcription and Alignment Auditor to review AI training audio in Italian. You will compare human or agent speech with annotator-produced transcriptions, assess word-level segment boundaries, apply structured error codes, and provide concise written rationales in English.
The role requires precise listening and consistent rubric application. You will identify critical errors first, distinguish genuine acoustic ambiguity from clear annotation mistakes, and issue a pass or fail judgment for each audit.
- Remote contract position
- Part-time schedule with a default commitment of 10 hours per week
- Pay: $39.50 per hour
- Entry-level experience classification
- Languages: Italian and English
- Audio-focused work involving transcription and evaluation
What You'll Do
You will review recordings of human or agent speech against existing transcriptions and verify that each word-level audio segment corresponds to the spoken content. Your written explanations should make the reasoning behind every judgment clear and concise.
- Check whether transcriptions accurately represent spoken Italian
- Verify that word-level timestamps match the actual start and end of each spoken word
- Confirm that reference text matches the spoken form
- Apply all relevant error codes
- Assign pass or fail verdicts using the required rubric
- Correct inaccurate segment boundaries when needed
- Split or merge segments when the alignment requires it
- Add missing word segments and remove segments containing no word
- Identify critical errors before secondary issues
- Explain audit decisions clearly in written English
Requirements
You should be a native or near-native Italian speaker, specifically familiar with Italian as spoken in Italy. Strong English reading and writing skills are required because audit conventions and decision rationales are handled in English.
Careful listening is central to the work. You must be able to distinguish similar sounds, identify word boundaries in continuous speech, apply written standards consistently across many judgments, and separate true acoustic ambiguity from clear transcription or alignment errors.
- Native or near-native Italian as spoken in Italy
- Strong English reading and writing skills
- Careful listening and accurate word-boundary judgment
- Ability to evaluate audio transcription and word-level timestamps
- Consistent application of error codes and pass or fail rubrics
- Ability to distinguish acoustic ambiguity from clear annotation errors
Helpful Background
Previous experience can be useful but is not required for this entry-level opportunity. Relevant backgrounds include transcription, subtitling, localization, linguistic annotation, and quality assurance.
Experience with AI data annotation or model evaluation, phonetics or linguistics, audio editing, or waveform tools may also help you work effectively in this role.
- Transcription or subtitling experience
- Localization or linguistic annotation experience
- Quality assurance or audit experience
- AI data annotation or model evaluation experience
- Background in phonetics or linguistics
- Familiarity with audio editing or waveform tools
Work Arrangement and Eligibility
This is remote contract work with a default workload of less than 20 hours per week and a planned commitment of 10 hours per week. The opportunity is available to contractors in the countries listed below.
- Austria, Belgium, Bulgaria, Canada, Cyprus, Czech Republic, Germany, Denmark, Estonia, Spain, Finland, France, United Kingdom
- Greece, Croatia, Hungary, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, Netherlands, Poland, Portugal
- Romania, Sweden, Slovenia, Slovakia, and the United States
Build Your AI Training Career
OpenTrain helps contributors discover fitting AI training work, build a credible profile, and develop a long-term portfolio of experience. Create a free OpenTrain account and apply for opportunities that match your language and audio-review skills.
- Work remotely on cutting-edge AI training projects
- Use your Italian language expertise in a specialized audio role
- Build documented experience in transcription auditing and data quality
- Apply through OpenTrain