How you fill your 52 weeks is your business. Helping you find your next opportunity is ours. Our Client’s solar design platform empowers users to create highly accurate solar system designs by combining best-in-class imagery, precise roof modeling, advanced energy production forecasting, and integrated financing workflows. Their internal design tool enables: - Accurate tracing of rooftops and rooftop obstructions - Automated panel placement using max-fit strategies - Energy production modeling across different orientations, shading conditions, and system sizes - Simulations for PV system performance, financial savings, and utility bill impacts You will lead the development of a Solar Design Copilot — an embedded AI agent that will: - Answer user questions about solar design, modeling decisions, PV system configuration, and production modeling best practices. - Explain the steps needed to create or modify a solar system design using our internal design SDK. - Execute user-directed design changes directly, modifying underlying design data structures through natural language prompts. This project lays the early foundation for an intelligent, agent-driven solar design experience, with the goal of making system design accessible, faster, and smarter while reducing support overhead. 1. Tech Stack & Tools We Use (and You’ll Likely Leverage) ⁃ Infrastructure: Google Cloud Platform (GCP), Cloud Functions - AI Platform: Vertex AI (for custom model deployments and managed endpoints) - Memory/Knowledge: Vector Databases (e.g., Pinecone, FAISS) for semantic retrieval - Orchestration: Building MCP (Model Context Protocol) servers for managing agent reasoning and task execution - Backend: Postgres (for design data storage), Python (primary language for orchestraAng AI logic) - Tools/Frameworks: LangChain or similar frameworks for chaining LLM tasks, Prompt engineering for design-specific workflows - Messaging & Events: Google Pub/Sub (for event-driven updates) You’ll have flexibility to propose and lead best practices for chaining, retrieval-augmented generation (RAG), memory persistence, and agent design patterns. 1. What We’re Looking For - Advanced experience creating, deploying, and scaling Agentic AI systems at enterprise-grade levels - 5+ years of experience in machine learning, AI engineering, or intelligent systems development - Proven ability to design and build autonomous or semi-autonomous agents that perform multi-step reasoning and execution - Strong experience working with vector databases, LLMs (OpenAI, Vertex AI, Claude, etc.), and AI orchestration frameworks (e.g., LangChain, AutoGen) - Solid backend engineering experience with Python, Postgres, and event-driven architectures - Deep understanding of retrieval-augmented generation (RAG), memory management, tool use planning, and safety/guardrails for AI agents - Familiarity with solar design workflows, PV systems, or a strong interest in learning technical domains quickly - Self-directed, experimental mindset — able to prototype, iterate, and scale solutions in a fast-paced environment Why 52 Limited: 52 Limited’s reputation in the recruiting industry is unparalleled in the Pacific Northwest. While we're headquartered right here in Portland, our tenured team routinely places world-class talent across the United States. We work with some of the most forward-thinking and recognizable clients in Portland, Seattle, Los Angeles, Austin, and beyond. To learn more about our team, check out www.52ltd.com/our/story. We are committed to fostering a people-first candidate experience that benefits our talent, clients, and community. 52 Limited encourages a diversity of applicants of all backgrounds and identities to apply. Let us know if you need reasonable accommodation during the application or interview process.