Senior Data Scientist – Information Retrieval & NLP
| Company | ZoomInfo |
|---|---|
| Location | San Francisco, CA, USA, Waltham, MA, USA, Vancouver, WA, USA, Bethesda, MD, USA |
| Salary | $167760 – $230670 |
| Type | Full-Time |
| Degrees | Master’s, PhD |
| Experience Level | Senior, Expert or higher |
Requirements
- 7+ yrs hands-on ML/NLP experience (or 4+ yrs post-PhD/Master’s) with at least two delivered, revenue-impacting products.
- Expertise in transformer stacks (BERT/GPT/T5), RAG, vector-based IR, and latency/throughput optimization.
- Proven track record building NER or entity-resolution systems at 100M+ record scale; knowledge-graph experience is a plus.
- Strong applied research chops (PyTorch or TensorFlow) paired with software-engineering rigor (Python, Go/Java a plus).
- Desire to work within MLOps tools and frameworks: Docker, K8s, GitOps, Terraform, feature stores, model registries, automated retraining.
- Ability to persuade exec and non-tech audiences with data-driven storytelling; comfortable owning strategy & budget.
Responsibilities
- Invent and productionize Transformer/RAG architectures that surface the right contact, company, or insight.
- Drive quantization, distillation, and SLM fine-tuning so models stay fast and affordable at petabyte scale.
- Prototype and launch hybrid dense/sparse retrieval pipelines on vector DBs (Pinecone, Weaviate, FAISS, OpenSearch).
- Own high-recall NER models that tag people, orgs, locations, and industry-specific entities across multi-language text.
- Build cross-dataset entity-resolution frameworks that dedupe hundreds of millions of records with sub-second latency; enrich with knowledge-graph signals where valuable.
- Design large-scale A/B and back-testing plans; close the loop from experiment to KPI uplift.
- Translate product goals into measurable ML KPIs; influence roadmap, capacity, and investment decisions.
- Mentor junior scientists/engineers; publish internal requirements documents, external blogs, and present at conferences.
Preferred Qualifications
- Knowledge-graph experience is a plus.