Principal Data Scientist

Wiley•Hoboken, NJ

1d•$140,000 - $200,733

About The Position

We are building systems that transform one of the world's largest scientific corporations into research intelligence. This involves production NLP pipelines processing millions of journal articles to extract entities, classifications, claim tuples, and summaries optimized for downstream agentic applications. We are seeking a principal data scientist to take end-to-end ownership of domain-specific content modeling work, from evaluation sets to pipeline deployment. This is a hands-on role within a small, senior team where data scientists own their models in production, including writing code, managing evaluations, shipping changes, and being accountable for outcomes. The ideal candidate will see their models through to real users in a dynamic market.

Requirements

Deep Python expertise, with experience writing it in production at scale, understanding tradeoffs between asyncio, threads, and queues.
Strong NLP background encompassing modern (LLMs, transformers, embeddings, retrieval) and classical (NER, classification, sequence labeling) approaches, with a history of building and learning from evaluations.
A habit of comparing approaches and selecting the most suitable one for a given task, capable of defending choices with evaluations and cost estimates, and addressing performance drift.
A proven track record of shipping systems that deliver value to real users, not just prototypes or papers.

Nice To Haves

Experience working with scientific or scholarly text.
Familiarity with AWS (S3, Batch, Lambda, SageMaker) and Parquet or Iceberg data lake patterns.
Experience running LLMs under real cost and latency budgets in production.
Some exposure to agentic AI applications, including tool use, multi-step reasoning, guardrails, and evaluation of trajectories.

Responsibilities

Design and build NLP enrichment pipelines that extract entities, classifications, claims, and summaries from scientific full-text at scale.
Compare NLP approaches to extraction and enrichment against LLM-based approaches, and select the most appropriate tool for each task, considering traditional NLP (NER, sequence labeling, classification), embedding-based retrieval, LLM prompting, and fine-tuned smaller models, defending choices with evaluation, cost, and operational tradeoffs.
Own evaluation by building golden sets in consultation with SMEs and vendors, choosing metrics, and making productive tradeoffs between speed, quality, and cost.
Write production-quality Python, manage concurrency and cost for high-volume LLM workloads, and structure code for engineerability and extensibility.
Collaborate with data engineers to orchestrate work in data pipeline and data build tools like Airflow and Dagster, designing idempotent, retryable, evaluable pipeline stages that remain reliable even when runs fail at scale.
Contribute to agentic AI application work, specifically tool-using systems that reason over the enriched corpus, shaping how agents ground and defend their answers with NLP and evaluation expertise.
Work directly with editors, product managers, and engineers, bringing a modeling perspective to product decisions and translating stakeholder feedback into concrete modeling work.