AI Data Scientist/Engineer (AWS Transcribe)
In Cyclad we work with top international IT companies to boost their potential in delivering outstanding, cutting-edge technologies that shape the world of the future. We are seeking an experienced Data Scientist/AI Engineer to join our customers, a leading company in consulting, technology services, and digital transformation.
Project information:
- Remote work: 100% remote
- Budget: 240 - 285 PLN/net/h + VAT
- Project: managing and improving systems that generate, summarize, and extract insights from audio and text data
- Seniority level:Senior
- Project language: English
- Start date: ASAP/ depending on candidate's availability
Project scope:
- Daily review of transcripts, summaries, data extracts, and task detections
- Growing Custom Language Model and Vocabulary for AWS Transcribe
- Catching edge cases, bugs, and LLMs hallucinations
- Improving prompts to address edge cases, bugs, and LLMs hallucinations
- Finding ways to improve transcripts (noise canceling, noise gates, de-noising, side chatter removal, new transcription models)
- Experiment with LLMs, parameters, and outputs
Requirements:
- Proficiency with AWS Transcribe for speech-to-text conversion
- Experience with Anthropic LLMs (Haikku and Sonnet) for summarization, task detection, and data extraction from documents (e.g., PDFs)
- Strong skills in propensity modeling, clustering, and segmentation
- Ability to perform daily reviews of transcripts, summaries, and task detections
- Familiarity with debugging and addressing bugs in LLM models.
- Ability to manage and process unstructured data sources (text, audio)
- English: B2
We offer:
- Full-time job agreement based on B2B
- Private medical care with dental care (covering 70% of costs) + rehabilitation package. Family package option possible
- Multisport card (also for an accompanying person)
- Life insurance