cover_letter / art_apWA3Xp2Rp4

Dear Waymo Product Management Hiring Team, Waymo's mission — to be the world's most trusted driver — is one of the few technology bets where the stakes are genuinely measured in human lives. With over ten million rider-only trips completed and a footprint now spanning 1,400+ square miles across 11 cities, the engineering organization powering that growth faces a compounding challenge: how do you keep hundreds of autonomous-driving engineers building at the frontier when the AI tooling landscape changes faster than any team can track? That question is exactly the kind I have spent the last several years building answers to — first at Intuit, where I owned developer platform infrastructure at scale, and now as a founder shipping production AI systems from the ground up. **Technical and AI Foundation** My engineering foundation runs deep enough to be practitioner-level on the tools this role requires. In 2004 I hand-coded backpropagation through time in C++ for a protein structure prediction system that became a NeurIPS 2014 publication. In 2026 I rewrote that same system in PyTorch, scaling from 413 parameters to 8 billion — a 19-million-fold increase — with MLflow experiment tracking, Optuna hyperparameter optimization, and FastAPI serving across six Docker containers. That arc is not decorative; it means I can sit with Waymo's ML engineers and speak precisely about training pipelines, not just wave at them. More directly relevant to the DevAI mandate: I built OpenClaw, a multi-agent orchestration framework implementing a gateway protocol, subagent delegation, profile management, and session switching — coordinating AI agent workflows across multiple industry verticals inside a production React/TypeScript desktop application. I also built and shipped MCP servers exposing screen-capture tools to AI coding assistants, and integrated the Claude MCP SDK into a live product used by real customers. When the Waymo JD calls out Model Context Protocol experience as a preferred qualification, that is not a checkbox I am reaching for — it is infrastructure I have already shipped. On the evaluation and measurement side, I built aeval, a local-first model evaluation platform with five core eval types, bootstrap confidence intervals, Welch's t-test, Cohen's d effect size, and automated safety gates integrated into CI/CD pipelines. Establishing rigorous, data-driven frameworks to measure the ROI of AI integration is not a concept I need to learn; it is a system I have already instrumented. **The Bridge** My career has moved consistently toward one specific problem: how do you make the engineers building complex systems dramatically more productive without sacrificing the rigor those systems demand? At Intuit that meant owning developer platform infrastructure at a scale most PMs never touch. As a founder it has meant building the AI-native tooling myself. The Waymo DevAI & Agentic Workflows role is the natural convergence of both tracks — and the autonomous driving context raises the stakes in a way I find genuinely motivating. **Why This Role** What excites me most is the scope of the organizational transformation described: not just shipping a tool, but redefining how autonomous driving software is built by converting internal workflows into AI-first paradigms. Waymo's engineering org is working on one of the hardest software problems in existence — perception, prediction, and planning for a vehicle operating in an uncontrolled physical world — and the DevAI platform you build will either accelerate or constrain every team working on that problem. The responsibility to cut through hype, organize fragmented information, and keep engineers focused on actionable insights is exactly the kind of thought-leadership function I have exercised both as a staff PM educating a CTO on enterprise language strategy and as a founder who has had to make every tooling bet with limited resources and high accountability. **Selected Relevant Experience** - **OpenClaw multi-agent orchestration framework** — designed and implemented gateway protocol, subagent delegation, profile management, and session switching; coordinating AI agent workflows across real estate, insurance, health/dental, and financial-markets verticals inside a production application. - **MCP server development** — shipped MCP servers exposing screen-capture tools to AI coding assistants; integrated Claude MCP SDK into a live, signed/notarized cross-platform desktop product. - **RL Post-Training Workbench** — built a 3-phase workbench benchmarking GRPO/DPO across TRL, VeRL, OpenRLHF, and NeMo RL; implemented 12 RL algorithms with standardized throughput/memory/convergence benchmarking and GPU Docker passthrough. - **aeval evaluation platform** — built CI/CD-integrated model evaluation with statistical rigor (bootstrap CIs, Welch's t-test, Cohen's d), adversarial safety testing, and automated regression gates — directly analogous to the DX/SPACE measurement frameworks called out in the JD. - **Intuit ICE Developer Platform** — achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23; reduced developer onboarding from 2–3 weeks to minutes; scaled throughput from 6K to 50K TPS via rSocket migration supporting ~1.5M concurrent connections at sub-25ms TP99. - **Enterprise-wide Service Language Assessment** — conducted analysis across 9 languages, synthesizing usage data and developer feedback into strategic investment recommendations presented to the CTO; directly analogous to the thought-leadership and education mandate in this role. - **Intuit SDK Starter Kits** — extended Java and Python SDKs with scaffolding templates, build configurations, testing frameworks, and CI/CD integration, enabling developers to go from zero to production-ready microservice in minutes. - **De Anza College — Adjunct Faculty** — teaching Fundamentals of Large-Scale Cloud Computing, Data Analytics, and Java Programming since 2018; demonstrated ability to translate complex technical concepts into structured education programs at scale. **Closing** Waymo's mission is not a metaphor — every improvement in developer velocity for the engineers building the Waymo Driver has a downstream effect on how quickly safer autonomous mobility reaches more people in more cities. I want to build the platform that makes those engineers faster, more confident, and better equipped to work alongside AI systems that are themselves rapidly evolving. I would welcome the opportunity to discuss how my background maps to the specific challenges your engineering organization is navigating. Thank you for your consideration. Sincerely, **O. Felix Amoruwa** famoruwa@berkeley.edu · 909-731-9011 · felixamoruwa.info