• Architect multi-agent systems using advanced orchestration frameworks (LangGraph, Google ADK) to automate complex customer support procedures end-to-end. • Build and scale integrations using Model Context Protocol (MCP) to connect LLMs with internal Coinbase APIs, databases, and third-party tooling. • Develop automated "LLM-as-a-judge" evaluation pipelines to monitor, measure, and improve the performance of non-deterministic AI agents in production. • Implement RAG, fine-tuning, and prompt engineering techniques to ensure chatbot responses are grounded, accurate, and compliant with Coinbase policies. • Ship production-ready Python services that are resilient, low-latency, and capable of handling Coinbase-scale traffic across asynchronous microservices. • Partner with Conversation Design and Product to translate complex business logic into executable agent procedures within the decentralized architecture.

• Develop AI algorithms, prototypes and solutions for healthcare, with a focus on foundation models and self-supervised learning; • Optimise models and training pipelines for accuracy, scale and rapid experimentation; • Follow agile methodology and software engineering best practice, focussing on test-driven development, rapid prototyping, validation and iteration; • Provide regular technical and other progress reports relevant to projects, and ensure all progression is properly documented; • Engage with the literature to benchmark against and adopt state-of-the-art techniques and algorithms; • Rigorously evaluate generative AI models, and partner closely with teams training models at scale; • Contribute to a culture of excellence, helping to solve problems as they arise, instil a culture of best practice, integrity and agility, as well as champion the Harrison mission internally and externally.

• Implement core deep-learning, computer vision, and (inverse-)procedural modeling algorithms in Python • Apply cutting-edge research in machine learning and computer graphics to solve real-world problems • Work closely with cofounders to translate customer requirements into technical milestones • Interact with remote machines via a Unix shell to deploy and test code on large-scale geospatial datasets • Use Git to manage source code and modularize complex implementation tasks

• Design and build training pipelines, data workflows, and model integration systems • Develop infrastructure that accelerates research iteration and reduces turnaround time • Build systems for data collection, curation, and preprocessing at scale • Create tools and automation that move experiments toward production readiness • Optimize data pipelines for reliability, performance, and observability • Collaborate with ML researchers to understand their needs and remove technical blockers • Work on model serving infrastructure and integration with the production framework • Write clean, well-tested code that maintains high engineering standards • Participate in code reviews and help raise the engineering bar across the team • Contribute to shared tools, infrastructure, and cross-role projects (20% Time) • Work with the dual-leadership model (Engineering Manager and Tech Lead) to understand priorities and technical direction • Document systems and decisions to support team knowledge sharing