Location: Bratislava (hybrid)
Contract: Freelance/FTE
Start: November 2025
Scope of Services:
- Design, fine-tune, and deploy LLM-based solutions for tasks like retrieval-augmented generation, question answering, summarization, NER, information extraction, prompt engineering, and zero/few-shot classification.
- Build pipelines for data preprocessing, embedding generation, vector indexing, and LLM inference.
- Process and analyze voice call transcripts and chat logs including speech-to-text cleaning, conversation segmentation, intent detection, multi-turn context handling, sentiment analysis, and automated report generation.
- Optimize models for latency, cost, and accuracy; package and deploy models and APIs.
- Collaborate with the Telco client’s team of data scientists and ML engineers; stay current on advances in LLM techniques.
Requirements for the supplier
- 2–3+ years experience delivering NLP and LLM solutions.
- Strong Python skills with experience in Hugging Face Transformers, LangChain, PyTorch/TensorFlow, and vector search tools.
- Solid understanding of transformer models and prompt engineering.
- Experience working with noisy text data, including ASR transcripts from calls and chats.
- Knowledge of MLOps practices for model serving and monitoring is a plus.
- Telecom experience is a plus.
- Experience with cloud-based platforms (GCP) is a plus.
