cover_letter / art_C4V6PUeeJhY

Dear Five9 Hiring Team, Five9 sits at a genuinely consequential inflection point: the shift from contact center software that assists agents to AI systems that *act* autonomously on behalf of customers across every touchpoint. That transformation — from passive tooling to agentic CX — is exactly the problem space I have been building toward, from hand-coding backpropagation through time in C++ at UC Berkeley in 2004 to architecting multi-agent orchestration frameworks and benchmarking RL post-training pipelines today. The GECX Principal PM role is the specific intersection of enterprise AI product leadership, agentic design, and platform-scale execution where my background is most directly applicable. --- **Technical and AI Foundation** My AI/ML work is hands-on and production-grade, not theoretical. In 2025–2026 I built an RL post-training workbench covering the full RLHF/DPO pipeline: a Reward Lab for designing and A/B testing reward functions across GSM8K, MATH, HumanEval, and UltraFeedback datasets; a Playground running real TRL-powered GRPO and DPO training with live SSE metric streaming on Apple Silicon and CUDA; and an Arena for head-to-head framework benchmarking across TRL, VeRL, OpenRLHF, and NeMo RL with GPU passthrough in Docker containers. I implemented 12 RL algorithms — PPO, GRPO, DAPO, REINFORCE, REINFORCE++, RLOO, DPO, SimPO, IPO, KTO, ORPO, and SPPO — with standardized throughput, memory, and convergence benchmarking across frameworks. This is the kind of depth that lets me write precise product requirements for LLM-powered features, evaluate vendor claims critically, and partner credibly with Google R&D on GECX integration. On the agentic architecture side, I built OpenClaw — a multi-agent orchestration framework with a gateway protocol, subagent delegation, profile management, and session switching — enabling coordinated AI agent workflows across distinct vertical domains. I also architected a RAG retrieval pipeline with ChromaDB vector store, multi-provider LLM orchestration (Claude, GPT-4, Gemini) with fallback routing, structured output validation, and token budget optimization. These are not integrations of existing platforms; they are ground-up implementations that required working through the same agentic design patterns — ReAct, tool use, multi-agent coordination — that the GECX integration will demand. For model evaluation, I built aeval: a local-first evaluation platform covering factuality, reasoning, instruction-following, safety, and code generation, with adversarial safety testing, refusal detection, bootstrap confidence intervals, Welch's t-test, Cohen's d effect size, and automated safety gates integrated into CI/CD. Responsible AI and rigorous evaluation are not afterthoughts in my process — they are first-class product requirements. My NeurIPS 2014 publication on artificial neural networks for protein secondary structure prediction, and the 2026 rewrite of that system spanning 413 parameters to 8B (a 19-million-fold scale increase), reflect a research foundation that predates the current LLM wave and gives me durable intuition about model behavior, training dynamics, and evaluation design. --- **Bridge** The through-line in my career is building developer-facing platforms and AI systems at the intersection of infrastructure scale and product experience — and the GECX role asks for exactly that combination: deep technical fluency in agentic AI, enterprise platform product leadership, and the ability to drive a strategic partnership with a hyperscaler while delivering measurable CX outcomes for Five9's customers. --- **Why This Role** The specific scope of GECX — Agent Assist, Conversational Insights, and Quality AI — maps directly to problems I find technically interesting and commercially important. Autonomous agents that execute multi-step actions across consumer touchpoints require the same orchestration primitives I built in OpenClaw, but operating at enterprise contact center scale with quality frameworks, compliance constraints, and real-time latency requirements. The opportunity to define Five9's integration strategy with Google's AI stack, represent that strategy externally, and build the evaluation and feedback loops that make agentic CX reliable is the kind of 0-to-1 product leadership challenge I am most effective at. --- **Selected Prior Experience** - **Intuit — ICE Platform (Staff PM):** Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma; scaled throughput from 6K to 50K TPS via rSocket migration supporting ~1.5M concurrent connections with sub-25ms TP99. Delivered ICE Self-Service platform (DevPortal, GitOps config, ICE Playground), reducing developer onboarding from 2–3 weeks to minutes in pre-prod and under 24 hours for production, while mitigating $1M+ in projected opex growth. - **Intuit — Developer SDKs:** Extended Java and Python SDK Starter Kits with scaffolding templates, build configurations, testing frameworks, and CI/CD integration — empowering developers to go from zero to production-ready microservice in minutes. This is directly analogous to building the integration surface and developer experience for Five9's GECX partnership. - **OpenClaw Multi-Agent Framework (StreamIO):** Implemented multi-agent orchestration with gateway protocol, subagent delegation, profile management, and session switching — enabling coordinated AI agent workflows across verticals. Directly applicable to designing autonomous agent capabilities within GECX. - **RAG + LLM Orchestration (Fintellect AI):** Architected RAG retrieval pipeline with ChromaDB, multi-provider LLM orchestration (Claude, GPT-4, Gemini) with fallback routing, structured output validation, and token budget optimization — the same architectural patterns required for Conversational Insights and Quality AI within GECX. - **aeval — AI Model Evaluation Platform:** Built evaluation platform with adversarial safety testing, statistical rigor (bootstrap CIs, effect size), and automated safety gates in CI/CD — directly applicable to championing responsible AI principles and evaluation frameworks across Five9's product organization. - **Splunk — Search Orchestration (Senior PM):** Owned Search Service (Go microservices), Search Catalog, and SPL/SPL2; delivered Scheduler Service end-to-end in ~4 months; led query performance optimization achieving up to 10x improvements for enterprise beta customers. Demonstrated ability to drive complex technical product delivery under timeline pressure with Fortune 500 customers. - **Kaiser Permanente — Logging-as-a-Service (Technical PM):** Led development and enterprise rollout of Splunk Logging-as-a-Service at 1.7 TB daily volume across 200+ internal enterprise customers, and built caching capability using Redis and XC10 for scalability and fault tolerance. Relevant to Quality AI data pipelines and the operational rigor Five9's enterprise customers require. --- **Closing** Five9's mission — bringing joy to customer experience — is not a soft aspiration; it is an engineering and product challenge that requires making AI agents reliable, explainable, and genuinely useful in high-stakes customer interactions. I have spent the last two years building the technical foundations — agentic orchestration, RL post-training, rigorous evaluation — that make that reliability possible, and the prior twelve years building the platform and enterprise product leadership experience to deliver it at scale. I would welcome the opportunity to discuss how my background maps to Five9's GECX roadmap. Sincerely, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info