cover_letter / art_LWcq2lgPP6g

Dear Spring Health Hiring Team, Spring Health's mission — eliminating every barrier to mental health — addresses one of the most consequential access problems of our time. The combination of clinically validated precision matching with agentic, AI-assisted servicing is exactly the kind of infrastructure-meets-impact work that drew me to platform product management in the first place. My path from hand-coding neural networks in C++ at UC Berkeley to building multi-agent orchestration frameworks and RL post-training workbenches has consistently been about one thing: making complex AI capabilities usable by the humans who need them most. **Technical and AI/ML Foundation** My AI/ML work spans both research and production. At NeurIPS 2014, I published on artificial neural networks for protein secondary structure prediction — work that originated from a hand-coded BPTT implementation in C++ and was later rewritten as a full PyTorch platform (BRAIN) spanning 413 to 8 billion parameters across five architectures, with MLflow experiment tracking, Optuna hyperparameter optimization, and FastAPI serving. More recently, I built aeval, a local-first model evaluation platform with five core eval types (factuality, reasoning, instruction-following, safety, code generation), adversarial safety testing with refusal detection, bootstrap confidence intervals, Welch's t-test, and Cohen's d effect size — integrated with CI/CD pipelines for automated regression detection and safety gates. My RL Workbench covers the full post-training pipeline: Reward Lab for designing and A/B testing reward functions across GSM8K, MATH, HumanEval, and UltraFeedback; a live TRL-powered GRPO/DPO training environment with SSE metric streaming; and an Arena for head-to-head benchmarking of TRL, VeRL, OpenRLHF, and NeMo RL across 12 algorithms including PPO, GRPO, DAPO, DPO, and SimPO. On the multi-agent 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 real estate, insurance, health, and financial services verticals. This is directly analogous to the agentic servicing architecture Spring Health is building to help human agents work at the top of their licensure. **Why This Role** The Lead PM, AI Servicing role sits at the intersection of operational efficiency, human-agent augmentation, and measurable clinical outcomes — a combination I find genuinely compelling. The specific success metrics in the JD (COGS/Gross Margin impact through agentic containment, CSAT, and time-to-resolution for member and provider issues) are the kind of outcome-oriented targets I've driven before, and the challenge of building tools that let human agents do more meaningful work — rather than replacing them — is a design problem I care about deeply. What excites me most is the scope: building AI-assisted workflows that streamline service delivery while preserving the human judgment that mental health support requires. That balance — automation where it helps, human expertise where it matters — is the same tension I navigated at Intuit when building the ICE Self-Service platform, and it's the right frame for regulated, high-stakes servicing environments. **Selected Prior Experience Most Relevant to This Role** - **Intuit — ICE Self-Service Platform:** Delivered a self-service developer platform (DevPortal, GitOps config, ICE Playground) that reduced onboarding from 2–3 weeks to minutes in pre-prod and under 24 hours for production, while mitigating over $1M in projected opex growth — a direct analog to driving COGS containment through agentic tooling. - **Intuit — ICE Engagement Scale:** 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 approximately 1.5M concurrent connections with sub-25ms TP99. - **Intuit — Telemetry-Driven Prioritization:** Used SQL and BigQuery to analyze usage data across approximately 20 mobile apps and 30+ product SKUs to prioritize developer pain points — the same data-informed, hypothesis-driven approach the JD calls out for measuring servicing impact. - **Fintellect AI — RAG and Multi-Provider LLM Orchestration:** 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 — directly applicable to building reliable, cost-efficient AI servicing workflows. - **StreamIO AI — OpenClaw Multi-Agent Orchestration:** Implemented a multi-agent framework with gateway protocol and subagent delegation enabling coordinated AI workflows across health, insurance, real estate, and financial services — the architectural pattern underlying agentic member and provider servicing. - **Splunk — Search Orchestration and Performance:** Owned Search Service (Go microservices), Search Catalog (PostgreSQL), and SPL/SPL2; led a query performance optimization initiative achieving up to 10x improvements for a beta enterprise customer — demonstrating the ability to drive measurable efficiency gains in complex, data-intensive platforms. - **Kaiser Permanente — Enterprise Platform at Scale:** Led development and enterprise rollout of Splunk Logging-as-a-Service (1.7 TB daily volume, 200+ internal enterprise customers) and built caching capability using Redis and XC10 to address scalability, fault tolerance, and data redundancy — experience operating in regulated, high-availability environments with demanding internal stakeholders. **Closing** Spring Health's mission to make mental healthcare accessible to everyone, everywhere is one I want to contribute to directly. The opportunity to build AI-assisted servicing tools that reduce friction for members in need — while enabling human agents to focus on the clinical and emotional work only they can do — is exactly the kind of product challenge that combines technical depth with meaningful human impact. I would welcome the opportunity to discuss how my background maps to what you're building. Warm regards, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info