We're looking for a Data Engineer with 4+ years of experience to own the caselaw and docket data layer at a seed-stage legal AI startup. You'll build production pipelines that ingest, normalize, and enrich data from PACER, NYSCEF, state court systems, and published opinions, turning raw legal data into the structured signal that feeds our AI strategy layer. Every downstream product surface depends on what you build. Legal data is uniquely messy, and prior fluency with it will save months of ramp.
Responsibilities:
Take full ownership of the caselaw and docket data layer: build and maintain production pipelines ingesting from PACER, NYSCEF, state courts, and published opinions at scale across millions of court records.
Wrangle messy semi-structured inputs (PDFs, scanned filings, XML/HTML) into clean, queryable structures.
Design LLM-assisted extraction workflows using Claude, Gemini, or OpenAI to turn unstructured legal text into reliable structured signal.
Work closely with full-stack engineers to feed judicial intelligence and AI strategy layers.
Run statistical analyses, optimize SQL queries, and leverage RAG and semantic search tooling (Pinecone, Voyage AI) to maximize data impact.
Operate and optimize the AWS data stack (S3, RDS) in line with SOC 2 subservice architecture guardrails.
This is a full-time, remote US-based role with a strong preference for NYC and Eastern Time overlap (4+ hours daily). Open to visa transfers (e.g. OPT, H1B transfers).
About the company: We are a small, high-intensity founding team building AI-powered tools for litigation and legal workflows. Engineers operate with high ownership and autonomy on production systems that demand correctness at scale. The team is primarily NY/NJ-based, with a preference for NYC and Eastern Time overlap, and gathers in person roughly once per quarter.
- 4-8 years of experience building production data pipelines
- Experience with large, dirty datasets: ability to understand, clean, and extract signal fast
- Strong AI background for data including RAG, semantic search, and LLM-assisted extraction
- Experience with legal data: PACER, dockets, court records
- Experience with regulatory or government data sources
- Bachelor's in CS, Data Science, or related field
- Proficient in Python and SQL for pipeline engineering
- PDF parsing, OCR, and document extraction tooling
- Vector databases or semantic search (Pinecone, Voyage AI)
- Fully autonomous; thrives without heavy management direction
- Deep curiosity about messy domain-specific data problems
- US-based with 4+ hours daily ET overlap
- Willing to travel to NYC for quarterly onsites
- Equity: 0.15% - 0.25%
- Open to visa transfers (e.g. OPT, H1B transfers)
- Remote US-based with strong NYC preference