Audio Transcription & Image Labeling for Autonomous Systems
Contributed to a dual-modality project involving both audio transcription and image annotation for an autonomous driving client. On the audio side, transcribed spoken instructions recorded in varied acoustic environments, applying speaker labels and time-stamps to build a high-quality voice recognition training dataset. On the image side, performed precise bounding box annotation on dashcam footage, labeling vehicles, pedestrians, traffic signs, and road markings to support the development of an autonomous driving model. Maintained strict adherence to client annotation guidelines throughout the project, delivering consistent and accurate output within tight turnaround windows. All work underwent quality review, achieving an accuracy benchmark above 96%.