cover_letter / art_2PxhmIiX8rs

Dear MongoDB Hiring Team, MongoDB has redefined what a database can be — not just a data store, but the foundational layer for how enterprises build, scale, and now reason with AI. The mission of empowering innovators to transform industries with software is one I have been living from the inside out: first as a Staff PM at Intuit scaling developer infrastructure to 675M+ annual engagements, and now as a founder building production AI systems from the ground up. When I read about this Staff PM – Internal AI role, I recognized the exact intersection where my technical depth, enterprise platform experience, and hands-on AI product work converge. **Technical and AI Foundation** My AI work is not advisory — it is built and shipped. In 2026, I completed 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/DPO training with live SSE metric streaming on Apple Silicon (MPS) 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, DPO, SimPO, KTO, and others) with standardized throughput, memory, and convergence benchmarking — the kind of rigorous evaluation infrastructure that separates production AI from demos. On the evaluation side, 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 — with CI/CD integration and automated safety gates. This is the responsible AI and governance layer that enterprise AI deployments require, built from first principles rather than inherited from a vendor checkbox. For multi-agent orchestration, I designed and shipped the OpenClaw framework inside StreamIO: a gateway protocol with subagent delegation, profile management, and session switching that coordinates AI agent workflows across real estate, insurance, health, and financial markets verticals. I also architected a RAG retrieval pipeline with ChromaDB, multi-provider LLM orchestration across Claude, GPT-4, and Gemini with fallback routing, structured output validation, and token budget optimization — directly analogous to the AI copilot and agentic platform work MongoDB is prioritizing internally. This technical foundation traces back further: my NeurIPS 2014 paper on neural networks for protein secondary structure prediction, and the original 2004 system I hand-coded in C++ with custom backpropagation through time — a lineage that gives me genuine intuition for where AI systems succeed and where they quietly fail. **Why This Role, Why Now** Enterprise internal AI is at an inflection point that mirrors what I navigated at Intuit: the gap between what AI can theoretically do and what actually gets adopted at scale inside a large organization. I closed that gap at Intuit by reducing developer onboarding from 2–3 weeks to minutes, scaling platform throughput from 6K to 50K TPS, and driving 275% YoY growth in platform engagements. MongoDB is at a similar moment — the database infrastructure is world-class, and the opportunity now is to turn that same rigor inward, transforming IT service delivery, knowledge management, and decision support with AI that actually works in production. **Role-Specific Connection** The JD calls out agentic platforms for organizations, bringing best practices and pitfalls from real deployments — that is precisely what I have been building and documenting. I have shipped multi-agent orchestration, evaluated LLM outputs under adversarial conditions, and managed the governance layer (responsible AI practices, safety gates, compliance-aware routing) that enterprise deployments require. The balance between quick wins like AI copilots and longer-term automation initiatives is a prioritization problem I have solved repeatedly, and I am drawn to MongoDB's scale — 60,000+ customers, 75% of the Fortune 100 — as the environment where that work has maximum leverage. **Selected Relevant Experience** - **ICE Self-Service Platform (Intuit):** Delivered DevPortal, GitOps config, and 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 — a direct analog to AI adoption and efficiency mandates in this role. - **675M+ Platform Engagements (Intuit):** Achieved 275% YoY growth in ICE engagements 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. - **OpenClaw Multi-Agent Orchestration (StreamIO):** Designed and shipped gateway protocol with subagent delegation, profile management, and session switching — enabling coordinated AI agent workflows across multiple industry verticals. - **aeval Evaluation Platform:** Built production AI evaluation infrastructure with adversarial safety testing, statistical rigor (bootstrap CIs, Welch's t-test, Cohen's d), CI/CD regression detection, and automated safety gates — the governance and responsible AI layer enterprise deployments require. - **RAG Pipeline (Fintellect AI):** Architected retrieval pipeline with ChromaDB, multi-provider LLM orchestration (Claude, GPT-4, Gemini) with fallback routing, structured output validation, and token budget optimization. - **Enterprise Service Language Assessment (Intuit):** Conducted enterprise-wide analysis across 9 languages with usage data and developer feedback, informing strategic investment decisions presented to the CTO — the kind of cross-functional stakeholder work and executive communication this role demands. - **Splunk Search Orchestration:** Owned Go microservices, PostgreSQL metadata service, and SPL/SPL2 roadmap; delivered Scheduler Service end-to-end in ~4 months and achieved up to 10x query performance improvements for a Fortune 500 beta customer. - **ICE Presence in Async Chat (Intuit):** Implemented feature generating $480K/month in additional invoicing — demonstrating the ability to connect platform work directly to measurable business outcomes. **Closing** MongoDB's stated purpose — enabling innovators to create, transform, and disrupt industries with software — is not a tagline I am reading for the first time in a job description. It describes the work I have been doing for over a decade, from enterprise platform infrastructure to production AI systems. I would welcome the opportunity to bring that experience to bear on MongoDB's internal AI strategy, and to help the organization that powers the AI era also lead it from the inside. Thank you for your consideration. **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info