Justin Waltrip

About

I’m Justin Waltrip, co-founder of Squary AI, legal AI that learns how a law firm produces its work and delivers that expertise consistently on every new matter. Squary helps mid-market firms produce firm-approved summaries, chronologies, and demand letters, with source-linked citations for every claim.

Based in Pittsburgh, I hold dual degrees in Computer Science and Applied Mathematics & Statistics from Johns Hopkins University, with a focus on machine learning.

Squary AI

Co-founder · Jan 2025 – Present

Squary is the expertise layer between a firm’s case management system and the underlying foundation models. It captures how a firm’s best attorneys frame issues and structure analysis, then applies that to every new matter, so work product stays consistent as models, documents, and staff change. Firms like Quatrini Law Group use Squary across personal injury, workers’ compensation, employment, and immigration practices.

  • Millions of pages processed for mid-market law firms
  • Document review up to 70% faster
  • 80%+ attorney adoption at a 40-attorney firm
  • 10x data volume at lower cost after consolidating search onto MongoDB Atlas (MongoDB customer story)

I lead engineering and product architecture, including:

  • A custom LLM orchestration framework that breaks large jobs into tracked steps, producing structured outputs at scale (e.g., 5,000+ row compliance matrices and due diligence classifications)
  • Document ingestion, indexing, and hybrid (keyword + vector) retrieval with page-level citations
  • Multi-tenant, security-first infrastructure with VPC isolation, encryption, and full audit trails
  • Integrations with SharePoint, Filevine, Google Drive, OneDrive, Dropbox, and Amazon S3

Cognistx

Data Scientist · May 2021 – Dec 2024

At Cognistx, an AI consulting firm, I delivered custom AI solutions for more than 10 clients across legal, healthcare, and retail.

  • Architected SQUARE, an enterprise question answering system combining extractive QA with retrieval augmented generation (RAG), which later became Squary AI
  • Designed a multi-tenant serverless AWS architecture and an evaluation framework for prompt engineering
  • Led development of graph neural network models for MoleculeAI, an AI drug discovery platform
  • Co-hosted the AI-Driven podcast on enterprise AI

Recognition & Speaking

In the News

Outside Work

Outside of work, I play ultimate frisbee and have recently taken up bouldering and pickleball. I also enjoy cooking and baking at home, and when time allows, backpacking, camping, waterskiing, and snowskiing.