Member of the Technical Staff

Norm•New York City, NY

2d•Hybrid

About The Position

Norm is hiring a Member of the Technical Staff to advance its multiplayer agentic harness, enabling people, agents, firm systems, and clients to collaborate on shared work with appropriate context, permissions, and accountability. The role involves designing multi-agent workflows, building reliable task and state management, improving agents’ ability to work across software, and capturing expert interactions as learning signals. It also includes developing memory and continual-learning systems, evaluation frameworks tied to economic outcomes, and collaborative interfaces for reviewing work and resolving exceptions. The position will help create tools for delegating and monitoring parallel work, requesting input, approving consequential actions, and giving firm leaders visibility into operations. It will also contribute to secure, high-stakes collaboration capabilities, legal-precedent tools, client portals, and demanding workflows such as due diligence, cap table reconciliation, and transactional document review. The role includes advancing Norm’s engineering practices through coding agents and agentic development, as well as conducting novel AI research grounded in professional work, publishing findings, and contributing to the research community. Norm prefers candidates who are practically fluent in AI and actively use it in their day-to-day work. Candidates should be prepared to discuss concrete AI tools, workflows, and outcomes during the interview process. The role is based in New York City with a hybrid work model.

Requirements

Practical, hands-on AI fluency and active use of AI in day-to-day work to support thinking, creation, and problem-solving.
Ability to discuss concrete examples of AI tools, workflows, and outcomes during the interview process.

Responsibilities

Advance the multiplayer agentic harness so humans, agents, firm systems, and clients can collaborate on shared work with appropriate context, permissions, and accountability.
Design multi-agent workflows using orchestration and actor models, enabling agents to divide and coordinate work and recover when plans change.
Build shared task and state management that reliably moves work between agents, attorneys, clients, and firm systems without losing context or ownership.
Improve agent effectiveness across software through computer use and deep integrations with professional-services systems.
Capture expert interactions, including delegation, corrections, reviews, and the context behind decisions, as learning signals.
Build memory and continual-learning systems that improve agent support for individuals and the firm and accommodate new training methods and frontier models.
Improve evaluation frameworks that measure agent performance against the quality, cost, time, and human effort involved in delivering client work.
Build dynamic collaborative interfaces for people and agents to review documents, compare alternatives, and resolve exceptions.
Develop ways to delegate work, monitor parallel work, request input, and approve consequential actions across teams and client relationships.
Give firm leaders a live view of operations, including priorities, ownership, dependencies, and progress across people, agents, and matters.
Build permissions, confidentiality controls, audit trails, secure execution, and reliable recovery into the platform.
Make legal precedent more useful to agents and apply general capabilities to workflows such as due diligence, cap table reconciliation, and transactional document review.
Build client portals and shared workspaces connecting client questions, documents, and decisions with attorneys and agents.
Advance engineering practices by building with coding agents, testing agentic-development approaches, and measuring their effects on delivery speed and software quality.
Conduct novel AI research grounded in professional work, publish findings, and contribute to the research community through collaboration and conference talks.