About the Role
We’re building high-quality training data from real Hindi-English telephone conversations, where speakers switch fluidly between both languages mid-sentence. As a Transcription & Annotation Specialist, you’ll produce accurate, verbatim transcripts and structured annotations that directly power AI speech systems. Precision, consistency, and a sharp ear for natural conversational speech are at the heart of this work.
Responsibilities
- Verbatim transcription — Transcribe Hindi-English telephone audio word-for-word, following our style guidelines closely (fillers, false starts, repetitions, and code-switching captured as spoken).
- Speaker labeling & event tagging — Mark speaker turns accurately and apply audio-event tags such as background noise, crosstalk/overlap, inaudible segments, silence, and non-speech sounds.
- PII redaction — Identify and redact personal information (names, phone numbers, account details, addresses, etc.) strictly per our PII handling rules to keep data compliant and secure.
- Volume & accuracy targets — Meet weekly throughput targets while consistently holding to our accuracy and quality benchmarks.
- Calibration & feedback — Take part in regular feedback and calibration sessions with the QA reviewer; apply corrections and evolving guideline updates to your ongoing work.
- Guideline adherence — Keep formatting, spelling, and tagging conventions consistent across every file, adapting quickly as the style guide is refined.
Requirements
- Bilingual fluency — Native or near-native Hindi and strong English proficiency; our calls code-switch between both languages mid-sentence, so you must follow and transcribe both effortlessly.
- Listening comprehension — Excellent ear for natural, fast, overlapping, conversational telephone speech, including accents, dialectal variation, and imperfect audio quality.
- Hinglish typing — Comfortable and consistent typing Hindi in Roman script (Hinglish) following our romanization conventions.
- Attention to detail — Meticulous and accuracy-first; you understand that correctness matters more than raw speed.
- Technical setup — Reliable computer, good-quality headphones, and a stable internet connection.
Nice to Have
- Prior experience in transcription, data annotation, subtitling, or linguistic QA.
- Familiarity with telephony/call-center audio or code-mixed (Hinglish) datasets.
- Experience working to style guides or SOPs in a remote, deadline-driven environment.
What We Offer
- Flexible, remote-first work with clear guidelines and QA support.
- Steady volume of work for reliable, high-accuracy contributors.
- The chance to shape real-world AI speech technology in Indian languages.
#LI-DNI