Senior Data Science Engineer (LLM.NLP)
- Рівень:
- senior
- Джерело:
- djinni.co
Що робити
- Conversational AI & LLMs: multi-turn dialogue, context handling and answer generation over financial data; prompt engineering, structured outputs, tool use and agent workflows; RAG pipelines over financial documents; fine-tune or train own LLMs where an API model isn't the best fit; reduce hallucinations, latency and cost per request.
- NLP models: train, fine-tune and maintain BERT-based models (intent detection, classification, entity extraction, embeddings); prepare and label data; decide when a smaller specialized model should replace an LLM call.
- LLM services in production: Python services within a 20+ microservice system; model routing, fallbacks, retries, caching, rate-limit handling; versioned REST APIs; clean, tested code and code reviews.
- Evaluation & monitoring: quality metrics and evaluation datasets; automated evaluation including LLM-as-a-judge and regression tests for prompt/model changes; tracing of LLM calls; experiments and A/B tests.
- AWS: train, deploy and operate models on AWS (EKS, Lambda, S3, SageMaker, GPU instances); CI/CD for models and prompts with DevOps; track and reduce cost.
Що пропонуємо
- RAG, vector databases and semantic search
- Fine-tuning or training LLMs
- Cloud platforms (AWS, GCP, Azure)
- LangChain, LangGraph, LlamaIndex or similar
- Observability: Grafana, Prometheus, Langfuse, LangSmith
Схожі вакансії
З блогу Trackr
Усі статті →Знайдено через trackr.help/jobs · Канал: @trackrhelp · Бот для персональних сповіщень: @trackrhelpBot


