AI & ML Engineer

Sandstone•New York City, NY

23h

About The Position

Sandstone is building the AI-native operating system for in-house legal teams. As an AI & ML Engineer, you will build the systems that make that possible: agents that reason over contracts and business context, retrieval systems that surface the right precedent, document understanding pipelines that turn legal work into structured intelligence, and evals that help us measure and improve quality. This is an applied AI product engineering role. You will own production AI systems end-to-end, from problem definition and data strategy through model behavior, orchestration, reliability, UX, and launch. You will work closely with engineers, lawyers, designers, and customers to translate legal judgment into product behavior. This is not a research role, a prompt engineering role, or a narrow ML infrastructure role. We are looking for product-minded AI engineers who can move quickly through ambiguity, work across the stack, and build systems that are useful, trustworthy, and delightful for lawyers to use every day.

Requirements

4+ years of experience as an AI engineer, ML engineer, full-stack engineer, product engineer, data scientist, or similar role
Experience shipping applied AI or LLM-powered products, especially involving retrieval, agents, evals, document understanding, structured extraction, or workflow orchestration
Strong software engineering fundamentals and comfort working across backend systems, data pipelines, APIs, and product-facing surfaces
Ability to own ambiguous technical and product problems from discovery through launch and iteration
Strong judgment around quality, latency, cost, reliability, observability, security, and UX tradeoffs
Excitement about working with lawyers and customers to encode domain expertise into product behavior
Bias toward shipping, measuring, and improving rather than waiting for perfect specs or perfect models
Clear communication, especially when explaining AI behavior and system tradeoffs to technical and non-technical teammates

Nice To Haves

Experience with legal, financial, enterprise SaaS, search, document AI, workflow automation, or other domains where correctness and trust matter

Responsibilities

Build production AI systems for legal workflows such as contract review, redlining, legal intake, drafting, negotiation, search, and knowledge management
Improve retrieval and context systems that reason over contracts, playbooks, precedent, customer data, and business context
Build document parsing and understanding pipelines that extract structure, entities, obligations, issues, and negotiation context from complex legal documents
Develop agentic and workflow-based LLM systems for long-running, multi-step legal tasks
Create evals, datasets, metrics, and feedback loops to measure quality, reliability, cost, and latency
Partner with lawyers and customers to turn real workflow feedback into product improvements
Work across backend AI services, data pipelines, APIs, and product surfaces
Make pragmatic system design decisions about when to use frontier models, smaller models, retrieval, rules, human review, or product design