Machine Learning Engineer, AI Agent Platform

The Company Arta is on an audacious and incredibly rewarding mission: to pave the way for people everywhere to lead more successful financial lives. Arta leverages AI and sophisticated digital tools—once reserved for ultra-high-net-worth individuals—and makes them accessible to a broader global audience. Think of it as your own digital family office, combining intelligent investment strategies, alternative assets, private market access, and smart automation to help you grow and protect your wealth effortlessly. We value trust, teamwork, and adaptability. Think: intelligent investing, personalized portfolios, and real-time trading, all backed by robust data infrastructure. The Role Arta is building the AI infrastructure for the next generation of wealth management. We partner with leading financial institutions to power strategic initiatives that create real competitive advantage, particularly in making high-quality, personalised advice scalable. Our platform enables intelligent agents to operate across core advisory workflows, from client servicing and suitability to portfolio research and analysis. These systems run in live, regulated environments and are embedded into how institutions serve their clients day to day. What You Will Do You will design and build production-ready agent systems that sit at the core of how financial decisions are supported and delivered. Build the AI Agent Platform Design and implement agent architectures (tool use, planning, memory, orchestration) Build systems for LLM orchestration , prompt management, and workflow execution Develop evaluation frameworks for agent quality, reliability, and safety Create benchmarking pipelines to measure model and system performance over time Enable Enterprise Deployment Build infrastructure for self-hosted and multi-tenant deployments Design systems that operate under enterprise constraints (security, latency, cost) Develop APIs and platform abstractions for external partners Bridge Research → Production Translate rapidly evolving LLM capabilities into stable, production-ready systems Partner with ML and product teams to integrate agents into real financial workflows Improve reliability, observability, and failure handling of agent systems Who You Are 5+ years building production ML systems or backend systems for ML-powered products Hands-on experience with LLMs, agent frameworks, or applied ML systems Strong Python skills and experience with modern ML tooling Experience with agent systems, tool use, or LLM orchestration frameworks Experience building evaluation / benchmarking systems for ML or LLMs Experience designing systems beyond notebooks — APIs, services, pipelines Strong systems thinking: latency, reliability, failure modes, tradeoffs Location: Bay Area preferred (3 days/week in our Mountain View office). We're flexible for exceptional remote candidates on the West Coast. Strong Plus Experience with self-hosted models or enterprise AI deployments Background in distributed systems or data infrastructure Exposure to financial systems or high-stakes domains What Makes This Role Different This is not a research or prototype-focused role. You will be responsible for shipping systems that operate in live financial environments. Your work directly supports institutional clients and real end users at some of the largest and fastest-growing financial institutions, not internal demos. If you’re motivated by making agent systems work reliably at scale, in complex and regulated settings, this role will be a strong fit. Interview Process Intro call with Recruiting Team, 30m - Virtual General Interview with Hiring Manager, 30m - Virtual Coding/Algorithm/Data Structure Interview, 60m - In Person or Virtual ML System Design Interview, 60m - In Person or Virtual AI Agent Coding Exercise, 120m - In Person or Virtual Culture/Co-Founder Interview, 30m - In Person or Virtual Note: We require at least one in-person interview before making our offer decision. For remotely locat