Senior NLP Engineer

Build NLP Systems to Support Real-World Legal Work

Legal1Up
is a small but growing legal technology company based in
Seattle, WA
with development operations based in
Guadalajara, Mexico
, providing document processing and litigation support tools to U.S. lawyers handling federal cases. We work at the intersection of AI, law, and structured data — going beyond simple drafting to build tools that support legal research, argument construction, and citation verification at scale.

We're looking for a
Senior NLP Engineer
to help us expand our in-house infrastructure for semantic search, legal fact extraction, and document analysis. You'll be part of a focused, hands-on engineering team (based in Mexico) working alongside legal professionals in Mexico and the Philippines. You'll build and optimize production-grade NLP systems that analyze complex court filings and legal exhibits — typically hundreds to thousands of pages long.

Key Responsibilities

• Build and optimize
  semantic search pipelines
  using vector databases (Pinecone) and transformer models
• Design NLP workflows for
  citation verification
  ,
  page-level fact extraction
  , and
  document-to-JSON structuring
• Contribute to tools that go beyond drafting — supporting
  legal research, fact pattern detection, and cross-document citation analysis
• Collaborate on the development of
  multi-agent AI systems
  that support specific litigation workflows (e.g., interrogatory generation, exhibit mapping)
• Integrate your models with existing APIs, storage systems (Azure, Google), and metadata (Airtable)
• Participate in design discussions, architecture reviews, and code deployment with the founder and lead developer

Requirements

• 5+ years of NLP experience
  working with transformer architectures in production settings
• Proficiency in the
  Python NLP ecosystem
  , including Hugging Face Transformers, Tokenizers, and Datasets
• Experience with
  vector search systems
  such as Pinecone, Qdrant, or Weaviate
• Comfortable working with
  long legal or technical documents
  (500–2000+ pages) and building pipelines with
  page-level granularity
• Familiar with
  FastAPI
  , RESTful APIs, and Azure-based infrastructure
• Able to balance model performance, cost, and reliability in real-world settings
• Must have
  professional working proficiency in English
  — the role involves reading and processing U.S. legal documents and business is conducted in English
• Able to work independently but communicate clearly and regularly to meet agreed deliverables

Nice to Have

• Experience in
  legal, regulatory, or document-heavy domains
• Familiarity with
  Retrieval-Augmented Generation (RAG)
  systems and prompt-based workflows
• Prior exposure to
  citation networks
  ,
  structured document outputs
  , or
  multi-agent architectures
• Background in working with
  legal datasets
  , OCR'd files, or structured annotation for fact verification

Work Structure

• Contract Role
  : Full-time commitment expected (not side project)
• Location
  : Remote; Mexico-based preferred
• Collaboration
  : Some
  occasional in-person meetings
  in Guadalajara with advance notice
• Work Hours
  : Flexible — you manage your own time, as long as you meet sprint goals and deliverables
• Team
  : Small development team + founder in Guadalajara, legal support staff in the Philippines and Mexico

Compensation

• $1,500 USD/month starting
  , negotiable based on experience
• Profit-sharing available
  based on team and company performance
• Long-term engagement preferred

About Legal1Up

We're building practical AI infrastructure for federal litigation support — not consumer legal tech or chatbots. Our customers are mostly
U.S. plaintiff-side law firms
, including those working in
employment, whistleblower, and civil rights cases
. We're focused on building reliable tooling that saves time and improves accuracy for real lawyers working on real cases.

We're at an early stage but growing quickly, with stable clients and a clear product direction. You'll be helping shape the technical core of a platform designed to handle millions of pages of evidence, build agent-based tools for discovery workflows, and enable lawyers to focus on strategy — not paperwork.