Principal Data Scientist

This is a fully remote position, open to applicants in New Jersey.

📋 Description

• Design and develop NLP enrichment pipelines that extract entities, classifications, claims, and summaries from scientific full texts on a large scale.

• Evaluate traditional NLP methods, embedding-based retrieval, LLM prompting, and fine-tuned smaller models, choosing strategies based on assessment, cost, and operational considerations.

• Create golden evaluation sets in collaboration with SMEs and vendors, select appropriate metrics, and balance speed, quality, and cost.

• Write production-level Python code and manage concurrency and expenses for high-volume LLM workloads.

• Structure code that is extensible, allowing engineers to implement it and data scientists to maintain it.

• Collaborate with data engineers utilizing Airflow and Dagster to orchestrate reliable, idempotent, retryable, and evaluable pipeline stages.

• Contribute to agentic AI applications that reason over the enriched scientific corpus.

• Partner with editors, product managers, and engineers to guide product decisions and translate stakeholder feedback into modeling tasks.

• Take ownership of domain-specific content modeling from evaluation sets all the way to production deployment.

• Manage models, evaluations, implemented changes, and production results.

⛳️ Requirements

• Extensive experience in Python production at scale for several years.

• Proficiency in asyncio, threads, and queues, including understanding their tradeoffs.

• Strong background in NLP, covering LLMs, transformers, embeddings, retrieval, NER, classification, and sequence labeling.

• Experience in building evaluations and learning from the results.

• Capability to compare modeling techniques using evaluations, cost estimates, and operational considerations.

• Proven track record of delivering production systems that provide value to actual users.

• Experience with scientific or scholarly text is preferred.

• Familiarity with AWS (S3, Batch, Lambda, SageMaker) and Parquet or Iceberg data lake patterns is preferred.

• Experience in running LLMs within production cost and latency constraints is preferred.

• Exposure to agentic AI applications, which include tool usage, multi-step reasoning, guardrails, and trajectory evaluation is preferred.

• A resume/CV is required when applying.

🏝️ Benefits

• Meeting-free Friday afternoons.

• Opportunities for professional development.

• Employee programs that support community, learning, and growth.

• Comprehensive benefits package.

• Competitive compensation.

• Reasonable accommodations for applicants and employees with disabilities.