The resource will deal with the following activities:

· Test and compare different Speech-to-Text systems on a set of business videos

· Definition of a pipeline: audio extraction →transcription → post-processing → validation

· Quality assessment with metrics and error analysis (reliability, terminology consistency, readability)

· Production of guidelines and best practices to make the process replicable

• Three-year degree in progress (Computer Engineering/Mathematics/AI/Data Science or related)
• Basic knowledge of Python
• Interest in NLP/Speech Recognition and experimentation on AI models
• Familiarity with Transformer models and libraries
• Experience (even basic) with STT tools (e.g. Whisper or equivalent)
• Interest in AI applications in the field of knowledge management (indexing, tagging, summarization)

‍