Engineering Lead, Agents

Whoever builds the leading intelligence for finance will create far more than returns. Poesis is the AI-native investment firm running autonomous agents that predict markets, construct portfolios, and manage risk. Our founders managed institutional capital at Capital Group ($3T AUM) and led enterprise ML at Goldman Sachs and Amazon. We're building a new type of firm, where live capital is the training ground for an intelligence that compounds with every signal.

At Poesis, we are pioneering AI for investing, opening the door to improved alpha discovery, higher quality decision-making and intelligent risk management. AI agents are a core part of how we make this a reality. We're looking for an experienced software engineer to help lead our agent engineering efforts. As the Agents Engineering Lead, you'll bring strong technical judgment and deep curiosity about what frontier models and harnesses make possible, translating emerging capabilities into production systems. You'll collaborate with Poesis leaders to set the vision and execute the development roadmap, with a unique opportunity to define the architecture that shapes how these systems are built and scaled across the organization.

• Provide technical leadership across the agent platform

• Design and build agent-driven workflows and internal systems

• Build eval, observability, and feedback systems that improve agent reliability, performance, and learning over time

• Establish patterns and infrastructure for reliable agent deployment

• 3+ years building software at a frontier AI lab, agentic startup, leading hedge fund, big tech company, or similar

• Strong familiarity with the modern AI stack (LLMs, tool use, orchestration, evaluation)

• Proficiency in Python and SQL, with bonus for TypeScript, Go, or Rust

• Basic understanding of financial markets and a personal or professional interest in investing

• Demonstrated experience designing, building, and deploying multi-agent systems in production environments; expertise in agent orchestration, tool calling, context window management, and agent grounding

• Proven ability to anticipate and address failure modes, with hands-on experience in observability and evaluation of agent systems

• BS/MS/PhD in Computer Science or related field, or equivalent practical experience
  

• Familiarity with agent safety, sandboxing, and guardrail design for high-stakes environments

• Experience working directly with researchers or quantitative developers as a technical partner

• Experience working with unstructured financial data (filings, earnings transcripts, news) and financial data providers (Bloomberg, CapitalIQ, Refinitiv)

Hybrid: 3 days per week on-site at our office in Menlo Park, CA. Relocation allowance available.

We offer excellent medical, dental, and vision coverage, alongside a strong benefits package that includes catered lunches in our Menlo Park office, commuter benefits, and more.
Current legal authorization to work in the US required; continuing work visa sponsorship available for full-time employees.

As an early team member, you’ll help shape not just the product, but how the company operates. Your decisions will have lasting impact across the business. You’ll build from first principles, with no legacy systems, or entrenched processes slowing you down. Our team is made up of people from elite companies and universities who are low ego, collaborative, and excited to build together.