Principal AI Architect

Overview

Mainspring Energy is revolutionizing power generation with the world’s most flexible and adaptable local power generation technology, the Mainspring Linear Generator. Commercial, industrial, and utility leaders are choosing Mainspring over traditional options like engines, turbines, and fuel cells to quickly and reliably deliver local power for EV charging, commercial facilities, data centers, and grid-scale operations.

The Mainspring Linear Generator is fuel flexible, ramps up and down to meet demand, and utilizes a flameless reaction with near-zero NOx emissions. Backed by top-tier investors including Khosla Ventures, Bill Gates, American Electric Power, Lightrock, and General Catalyst, Mainspring designs, manufactures and delivers its products to customers across the U.S. today, and we’re quickly scaling for international expansion.

Inspired by our vision of the affordable, reliable, net-zero carbon grid, Mainspring is rapidly expanding within the $816B global electricity equipment market, and we’re hiring the best talent to meet growing customer demand around the globe. We welcome a broad range of backgrounds, experiences, and talents to bring fresh perspectives and ongoing innovation to our customers.

More information can be found at

Mainspring Energy is evolving beyond using AI for simple productivity gains. We are seeking a Principal AI Architect to lead our transition into an organization where AI agents are fully embedded in our engineering development cycles. This is a strategic leadership role focused on architecting the future of how we build linear generator technology.

At the principal level, you will not just support existing processes; you will define the AI strategy for engineering development workflows, ensuring that AI-driven automation has clear structure, rigorous guardrails, and measurable Return on Investment (ROI).

AI Strategy for Engineering Workflows

• Architect Embedded Agent Cycles: Transition Mainspring from "AI-assisted" tasks to a model where autonomous and semi-autonomous AI agents are natively integrated into the engineering lifecycle—from code generation and hardware simulation to automated testing and documentation.
• Workflow Optimization: Identify and re-engineer high-impact engineering bottlenecks, deploying AI agents that act as functional members of the development team.
• ROI & Performance Measurement: Establish a data-driven framework to measure the impact of AI on engineering velocity, error reduction, and resource optimization; define and report on clear ROI metrics for all AI initiatives.

• Standardized Guardrails: Design and implement automated guardrails within the CI/CD pipeline to ensure AI-generated output meets safety, security, and quality standards.
• Responsible Agent Execution: Establish governance frameworks for agentic behavior, including "human-in-the-loop" checkpoints, audit logs, and access controls to prevent drift or unauthorized actions in production environments.
• Model Evaluation: Build rigorous benchmarking systems to evaluate model reliability and cost-effectiveness, ensuring a lean and high-performing AI stack.

• Agentic Infrastructure: Design the end-to-end infrastructure required to support multi-agent systems, including orchestration layers, long-term memory (Vector DBs), and tool-use capabilities (RAG).
• Build vs. Buy Strategy: Act as the technical authority on whether to leverage third-party AI platforms or develop proprietary agentic tools tailored to our specific hardware-software engineering needs.

• 12+ years of experience in software engineering or ML infrastructure, with a proven track record of moving AI beyond "chatbots" and into functional, automated workflows.
• Deep Expertise in Agentic Frameworks: Proven track record in building autonomous systems using the Gemini and Anthropic SDKs (specifically via Vertex AI). Expert-level proficiency in architecting agentic design patterns—including MCP (Model Context Protocol), multi-agent orchestration, and complex state management—within high-scale production environments.
• Stra…