Staff Data Scientist (NLP)
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
TL;DR
Staff Data Scientist (NLP): Building and optimizing NLP models to transform unstructured healthcare data into actionable product features with an accent on LLM fine-tuning, RAG pipelines, and agentic workflows. Focus on end-to-end ownership of ML problems, from framing and data preparation to production deployment and performance evaluation.
Location: Must be based in the US
Salary: $180,000–$230,000+
Company
is a healthcare intelligence company providing an AI-powered platform to improve financial outcomes and transparency for health plans.
What you will do
- Design and train NLP models for classification, entity extraction, retrieval, and summarization.
- Fine-tune LLMs and build RAG pipelines or agentic workflows.
- Develop robust evaluation harnesses including offline metrics and human-in-the-loop reviews.
- Partner with ML engineers to productionize models, focusing on latency, cost, and drift detection.
- Translate ambiguous product requirements into well-scoped ML problems for stakeholders.
Requirements
- 3+ years of applied ML experience with a focus on NLP.
- Strong proficiency in Python and modern NLP stacks like PyTorch, JAX, or Hugging Face.
- Hands-on experience with transformer fine-tuning (LoRA, instruction tuning) or RAG systems.
- Solid understanding of evaluation metrics for LLMs and NLP tasks.
- Experience with SQL, vector databases, and major cloud platforms (AWS/GCP/Azure).
- Must be based in the US.
Nice to have
- Publications at major AI conferences (ACL, EMNLP, NeurIPS).
- Experience with multilingual, speech, or multimodal models.
- Background in shipping LLM features within regulated industries like healthcare or finance.
Culture & Benefits
- Fully remote work environment within the US.
- Comprehensive medical, dental, and vision insurance.
- 401(k) plan with company match.
- Unlimited PTO policy.
- Tuition reimbursement and professional development support.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →