Design Engineer

Manifest OS•New York City, NY

9d•Onsite

About The Position

Manifest OS is an AI-native company on a mission to replace the billable hour and make legal services more accessible for American businesses and consumers. We power the next generation of AI-native law firms with one unified global brand, a proprietary technology platform, and a centralized back office — enabling lawyers to eliminate the administrative burden and focus on delivering exceptional outcomes for their clients. Manifest OS has raised a $60M Series A from Menlo Ventures, Kleiner Perkins, First Round, and Quiet Capital. About the Role We're looking for a Design Engineer to help build the legal tech platform powering the next generation of AI-native law firms. You care as much about how software feels as how it works. You'll own the development of secure client portals and enterprise-grade systems - from architecture to production. This opportunity is ideal for engineers who want broad scope, high ownership and move nimbly across lanes. You'll work across AI systems architecture and full-stack engineering - eval systems, retrieval, agent orchestration, and optimization, alongside the APIs, data models, and systems that power the core platform. This is an in-person role - you'll be on-site at our NYC office 5 days a week.

Requirements

7+ years building production software, with demonstrated depth in at least one area (backend systems, infra, ML/AI systems, frontend or similar)
Built or meaningfully shaped a production design system, including component APIs, tokens, theming, and documentation, that other engineers adopted and shipped with
Proven track record delivering features that produce measurable business outcomes
Comfort with relational database design and data modeling
Hands-on experience deploying and operating services on a major cloud platform
Fluent with AI coding tools as a core part of your daily workflow
A track record of self-teaching your way into unfamiliar systems and shipping there
First-principles and systems thinking: you can reason about a system you didn't build, from the database up through the API to the product surface
Familiarity with our stack - TypeScript, React, GCP, PostgreSQL

Nice To Haves

Experience making a design system legible to AI coding tools, through machine-readable documentation, token linting, or visual regression testing that keeps generated UI on-system
Deeply engaged with the AI tooling ecosystem - you follow new model releases, experiment with new tools as they land, and stay current out of genuine curiosity
Experience with eval systems, retrieval, agent orchestration OR ML in production
Background building client portals, CRMs, or case management systems
Familiarity with legal tech or other regulated industries

Responsibilities

Evolve the design system into scalable components and tokens that make every project fast to ship and beautiful by default, whether a person or an agent is building it.
Spot patterns that keep recurring across the platform, and promote them into tested, documented primitives that designers and engineers can compose without reinventing.
Build the guardrails that keep the product on-system as it scales, including token linting, visual regression testing, and component documentation that AI coding tools can read and follow.
Translate the design system into components and tokens that make the next screen fast to ship and beautiful by default, by a person or an agent
Partner with design on what AI-native interfaces should be in a high-stakes domain: progressive disclosure, streaming states, and the patterns nobody's figured out yet
Define the contracts between backend and UI - data fetching, caching, state boundaries, the choices that make the design hold up against real data
Own the screens clients and attorneys live in, where a well-set page turns a dreaded process into one that actually feels good to use
Build loading, empty, error, partial, and permission-restricted states as first-class parts of every feature, so nothing ships with only the happy path.
Treat accessibility, performance, and failure recovery as acceptance criteria, not polish. A failed document upload the night before a filing deadline should tell the user exactly what happened and what to do next.
Own end-to-end features across our legal tech platform - client portals, case management, and the workflows lawyers and clients depend on daily
Design production-grade APIs, data models, and permission systems for a regulated, high-trust domain
Eval systems, retrieval, and agent orchestration - building the infrastructure that measures and improves how our AI agents perform
Harness engineering - building the tooling and guardrails that let the rest of the company build securely, reliably, and self-serve