cover_letter / art_pCMJm-2ceh8

Dear SoFi Bank Platform Hiring Team, SoFi is doing something genuinely rare in financial services: building the full stack — the member-facing products *and* the infrastructure underneath them. The ambition to become the "AWS of Banking" is not a marketing line; it is a structural bet that whoever owns the rails owns the relationship. That framing resonates directly with my experience at Intuit, where I owned the shared infrastructure layer — ICE — that every product team across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma depended on to move data and trigger actions at scale. I know what it means to be the platform that everyone else builds on, and I know the discipline it requires. ## Technical and Platform Foundation At Intuit, I owned the ICE platform end-to-end as Staff Product Manager for Developer Frameworks and Platform Infrastructure. The numbers that matter for this role: I scaled ICE throughput from 6,000 to 50,000 transactions per second via an rSocket migration supporting approximately 1.5 million concurrent connections at sub-25ms TP99 latency. That is not a consumer feature — that is infrastructure engineering at the level SoFi's real-time payments platform requires. I also delivered the ICE Self-Service platform (DevPortal, GitOps config, ICE Playground), reducing developer onboarding from two to three weeks down to minutes in pre-production and under 24 hours for production, while mitigating over $1M in projected opex growth. Scaling to 675M+ engagements in FY23 required defining SLAs, monitoring platform health, and building operational tooling that support teams could act on — exactly the operational excellence mandate in this role. On the financial infrastructure side, I built Fintellect AI, a mobile-first investing and financial education platform where I architected a RAG retrieval pipeline with ChromaDB, multi-provider LLM orchestration across Claude, GPT-4, and Gemini with fallback routing, and structured output validation with token budget optimization. While Fintellect is an AI product rather than a payments rail, the architectural patterns — fallback routing, provider orchestration, transaction lifecycle management — are directly analogous to payment hub design. I understand how to build systems that must not fail, and how to design graceful degradation when a provider is unavailable. My technical depth extends to hands-on work: I can read an API spec, write SQL against BigQuery or PostgreSQL to investigate an anomaly, and review an architecture diagram credibly with engineers. At Splunk, I owned Go microservices (Search Service), a PostgreSQL metadata service (Search Catalog), and SPL/SPL2 query language — and I drove a query performance optimization initiative that achieved up to 10x improvements for a beta Fortune 500 customer. At Kaiser Permanente, I led enterprise rollout of a Redis-based caching capability addressing scalability, fault tolerance, and data redundancy across the SOA product suite, alongside Splunk Logging-as-a-Service at 1.7 TB daily volume serving 200+ internal enterprise customers. ## Why This Role The arc from shared infrastructure platform (Intuit ICE) → financial product infrastructure (Fintellect AI) → this role is a direct line. SoFi's real-time payments platform is precisely the intersection: shared infrastructure, financial services domain, and the complexity of serving multiple internal business lines — Consumer, Commercial, Invest, Lending — each with different latency, compliance, and cost requirements on the same rails. What specifically draws me to this role is the scope of the rail portfolio: FedNow, Fedwire, RTP, Visa Direct, Mastercard Send. Each rail has distinct message formats, settlement mechanics, cutoff windows, and regulatory obligations. Designing the routing logic, fallback strategies, and rail selection framework that sits above all of them — a payment orchestration layer — is a genuinely hard platform problem. I have built the analogous layer at Intuit (protocol selection, rSocket migration, throughput optimization) and I want to apply that pattern to money movement infrastructure where the stakes are higher and the domain is more consequential. ## Selected Relevant Experience - **ICE platform scale:** Scaled throughput from 6K to 50K TPS via rSocket migration; 675M+ engagements in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma; sub-25ms TP99 at ~1.5M concurrent connections. - **Developer platform and onboarding:** Delivered ICE Self-Service (DevPortal, GitOps, ICE Playground), reducing onboarding from 2–3 weeks to minutes in pre-prod and <24 hours for production; mitigated $1M+ in projected opex growth. - **Instrumentation and metrics:** Used SQL and BigQuery to prioritize developer pain points across ~20 mobile apps and 30+ product SKUs; built Asterias, a declarative asset lifecycle management platform with GraphQL API — establishing the data infrastructure to track what matters. - **Vendor and partner management:** Led Mailchimp GCP-to-AWS migration for MSaaS, coordinating across infrastructure vendors and internal partners to meet production deadlines; managed enterprise-wide Service Language Assessment across 9 languages presented to CTO. - **Operational tooling and drift detection:** Initiated MSaaS Drift Detection and Resolution program — wrote a Java JAR library to scan Git repos for configuration drift, partnered with Design on DevPortal UI, and built a remediation roadmap using OpenRewrite. - **Multi-provider orchestration with fallback routing:** Architected Fintellect AI's LLM orchestration layer with fallback routing across Claude, GPT-4, and Gemini — the same architectural discipline required for payment rail selection and fallback in a payment hub. - **Real-time infrastructure with latency SLAs:** Implemented ICE Presence in async chat (generating $480K/month in additional invoicing) and deployed Background-to-Foreground Messaging on iOS/Android with <100ms latency — demonstrating ownership of real-time infrastructure with hard latency requirements. - **Enterprise-scale logging and caching infrastructure:** At Kaiser Permanente, led Splunk Logging-as-a-Service (1.7 TB daily, 200+ customers) and Redis-based caching for fault tolerance and scalability — foundational infrastructure patterns directly applicable to payment platform observability and resilience. ## Closing SoFi's mission — helping members get their money right — is most credible when the infrastructure underneath it is invisible: instant, reliable, and always available. The real-time payments platform is where that credibility is built or lost, one transaction at a time. I have spent twelve years building the kind of shared infrastructure that other teams depend on, and I understand both the technical discipline and the internal evangelism required to make platform adoption succeed. I would welcome the opportunity to bring that experience to SoFi's Bank Platform team. Thank you for your consideration. --- **O. Felix Amoruwa** famoruwa@berkeley.edu · 909-731-9011 · felixamoruwa.info