Description
Ready to do the most impactful work of your career?
At Coinbase, we are uncompromising on our mission to increase economic freedom. The bar is high, the environment is intense, and we like it that way. This isn't a place for complacency, it’s a place to be pushed past your perceived limits. If you're ready to build the future of finance alongside people who refuse to settle for "good enough," you belong here.
Coinbase's Legal organization is at an inflection point. We manage corporate governance for 60+ global subsidiaries, public company reporting, equity programs, a venture portfolio, M&A, every institutional client agreement, and product counsel for all of Coinbase's PGs , from tokenized equities to stablecoin launches. Our clients are the CEO, COO, and every product group at the company. We work at the speed of a crypto startup with the compliance requirements of a public financial institution.
We are not looking to incrementally improve. We are looking to reimagine how legal work gets done at scale , replacing manual, high-volume workflows with AI-native systems, building an agent platform where today's junior lawyers become tomorrow's agent builders, and proving a model for how every non-technical org at Coinbase can transform with AI.
We have a roadmap, we have playbooks, and we have deep domain knowledge. What we need now is a technical architect to build the platform that brings this vision to life.
Responsibilities
- Architect and build Legal AI infrastructure , Design and ship a platform of AI agents and automated workflows that replace high-volume, templated legal work including institutional client agreement redlining, SEC filing support, subsidiary governance workflows, and commercial deal processing.
- Own the institutional client agreement automation system , Our highest-priority use case: build an end-to-end agentic workflow that ingests counterparty markups, evaluates them against a structured playbook, generates a first-pass redline with comments, routes for human review, and tracks outcomes. This is a production-grade system affecting 220+ institutional clients in pipeline.
- Build a cohesive agent platform and governance layer , Consolidate a fragmented set of ad-hoc bots and agents into a consistent, documented, maintainable platform with naming conventions, personas, observability, and standards for how Legal agents are built and deployed.
- Act as the technical thought partner to the VP and the broader Legal leadership team , Translate complex legal workflows into engineered systems, and translate engineering constraints back to non-technical stakeholders. The ability to bridge these worlds is as important as the engineering itself.
- Define the make-vs-buy strategy , Evaluate and retire third-party legal tech tools (Harvey, etc.) in favor of purpose-built internal solutions where Coinbase's engineering strength gives us a structural advantage.
- Be the case study , This role is explicitly designed to be a pilot for forward-deployed engineering in non-technical organizations at Coinbase. Success here creates the playbook for how other functions (Finance, HR, Operations) can undergo the same transformation.
Requirements
- 12+ years of backend software engineering experience, with a track record of shipping production systems at scale
- Proven experience getting LLM-based applications to production , including evaluation frameworks, grounding, guardrails against hallucination, and human-in-the-loop design
- Multi-agent orchestration experience , you've designed and operated systems where multiple AI agents collaborate, hand off work, and escalate to humans appropriately
- Exceptional ability to communicate with and learn from non-technical domain experts , you don't need to know legal terminology on day one, but you need to rapidly internalize complex workflows and translate them into engineered systems
- Platform mindset , you think in primitives, reusability, and standards, not just point solutions
- Distributed systems fundamentals , the agreement automation system will touch document processing, structured playbook evaluation, routing logic, and audit trails; you need to architect this correctly from day one
- Relentless drive to set and achieve ambitious goals in ambiguous, greenfield environments
- Demonstrates the ability to responsibly use generative AI tools and copilots (e.g., LibreChat, Gemini, Glean) in daily workflows, continuously learn as tools evolve, and apply human-in-the-loop practices to deliver business-ready outputs