NLP Research Engineer

About The Role The role bridges state-of-the-art language model research and the constraints of real production systems - latency, cost, interpretability, and compliance. You will work directly with research teams to fine-tune, evaluate, and deploy NLP models, and with engineering teams to make those models work reliably in client environments. Key Responsibilities • Fine-tune and evaluate transformer-based models (BERT, T5, LLaMA, Mistral) for named entity recognition, relation extraction, classification, and text generation tasks • Build systematic evaluation frameworks: benchmark datasets, human-evaluation pipelines, and automated regression suites for NLP model quality • Develop preprocessing pipelines for unstructured text - legal documents, clinical notes, financial filings - including OCR post-processing and entity normalization • Collaborate with product engineers to deploy NLP models within latency and cost constraints; implement distillation, quantization, and caching strategies as needed • Stay current on NLP/LLM research literature and evaluate its relevance to active client problems; bring relevant advances into production-ready implementations • Document model architecture, evaluation results, and known limitations clearly for both technical teammates and client-facing stakeholders • Participate in literature reviews, internal research discussions, and occasional external publications or conference presentations What We Are Looking For • 1–4 years of applied NLP or ML engineering experience, with hands-on model development beyond prompt engineering • Deep familiarity with Hugging Face Transformers, tokenizers, and the broader transformers ecosystem • Strong Python; experience processing and cleaning real-world unstructured text data at scale • Understanding of core NLP concepts: tokenization, embeddings, attention mechanisms, sequence labeling, span extraction • MS or PhD in Computer Science, Computational Linguistics, Statistics, or a related field strongly preferred • Ability to read and implement published NLP research papers independently • Bonus: experience with spaCy, Prodigy annotation tools, LLM-based NER, or RLHF methodologies Location Pittsburgh, PA (Carnegie Mellon corridor) • New York City • Boston • San Francisco • Seattle • Remote strongly considered