Speech Recognition Jobs

I have a batch of videos—each running under the 30-minute mark—that I need transcribed accurately and promptly. The files will be supplied in common formats (MP4 or similar), and I’ll need the spoken content delivered back as neatly formatted text ready for immediate use. When you respond, please focus on the depth of your transcription experience: the volume you’ve handled, any industry niches you’re comfortable with, and your typical turnaround for clips of this length. If you work with standard tools such as Express Scribe, Otter, or similar speech-to-text software before final human proofreading, feel free to mention that workflow. I value clean, well-punctuated transcripts with speaker changes clearly indicated and minimal need for follow-up edits. Let ...

I need a robust solution that can pick out individual human voices even when the recording is full of background clatter—think cafés, factory floors, or busy streets. The sole objective is to identify who is speaking; transcription and emotion analysis are outside the scope. I’m open to any modern approach—deep-learning architectures such as ECAPA-TDNN, x-vector, or a custom CNN-RNN hybrid—so long as the final model remains reliable when the signal-to-noise ratio drops. Python is preferred for the pipeline, and frameworks like PyTorch or TensorFlow are perfectly acceptable. Please work with publicly licensable datasets or clearly state any proprietary material you intend to use, and describe your noise-augmentation strategy up front so I can vet it. Deliver...

I have a batch of Amharic interviews from Project PW that I need converted into clean, verbatim text. Each recording runs between 30 and 60 minutes. The audio is generally understandable but there is occasional background chatter, so a good ear for speech buried under light noise will help. What I need from you: • A separate transcript for every file, typed in standard Amharic script and saved as .docx or .txt. • Verbatim capture—no translation or summarising. • Clear tagging of any inaudible sections (e.g. [inaudible 00:12:34]). The work is fully remote and confidential. Let me know your experience handling Amharic audio, the tools you like to use, and your typical turnaround time per hour of audio. I will send over a short paid sample first; once that is ap...

Train OpenWakeWord with a Custom Wake Word (High Accuracy, Low False Positives) Project Overview I am looking for an experienced AI/ML engineer to train OpenWakeWord with a custom wake word. The goal is to build a production-quality wake word model that is highly accurate, responsive, and resistant to false activations. Scope of Work * Train a custom OpenWakeWord model from scratch or fine-tune an existing pipeline. * Use real positive voice samples from multiple speakers, accents, speaking speeds, distances, and environments. * Collect and use a large negative dataset containing similar-sounding words, random conversations, TV/audio, background noise, music, and other speech to minimize false positives. * Optimize the model for: * High detection accuracy * Extremely low false ...

## **Native Malaysian English Speakers Needed – AI Voice Recording Project ($12/hour)** **TRS Multilingual** is looking for contributors with **fluent Malaysian English (Malay-accented English)** to join our AI voice data collection project. The task is simple: record natural, spontaneous speech in English with your natural Malaysian accent and upload the recordings through our platform. ### **About the Project** This project helps improve AI speech recognition and transcription systems. Participants will record natural spoken English using a **Malaysian accent**, covering everyday topics such as personal experiences, daily life, hobbies, travel, food, technology, and casual conversations. ### **Responsibilities** * Record natural spoken English with a **Malaysian accent** (10&...