cover_letter / art_M83V3eXmc6o

Dear SoFi Hiring Team, SoFi is doing something structurally distinct in financial services: building the full stack — banking charter, core infrastructure (Technisys), and B2B platform (Galileo) — while simultaneously serving 9M+ members through a mobile-first consumer experience. That integration of infrastructure depth with consumer reach is rare, and it creates the conditions where an AI-powered SDLC platform can compound across every product surface simultaneously. My path from hand-coding backpropagation through time in C++ at Berkeley in 2004, through NeurIPS-published ML research, to shipping production AI platforms and developer infrastructure at Intuit scale, has been building toward exactly this kind of role. ## Technical and AI Foundation My AI/ML work spans from foundational research to production systems. My 2014 NeurIPS paper on artificial neural networks for protein secondary structure prediction established early credibility in deep learning at a time when the field was nascent. In 2026, I rewrote that original C++ system into a production ML platform in PyTorch — five neural architectures (feedforward, GRU, Transformer, ESM-2, multi-task), MLflow experiment tracking, Optuna hyperparameter optimization, FastAPI serving, and 823 automated tests across six Docker containers — scaling from 413 to 8B parameters, a 19-million-fold increase. More directly relevant to this role: I built a full RL post-training workbench covering the complete RLHF/DPO pipeline. The platform includes a Reward Lab for designing and A/B testing reward functions across four datasets (GSM8K, MATH, HumanEval, UltraFeedback), a live training Playground powered by TRL with real-time 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 evaluation rigor the JD describes when it calls for measurable frameworks ensuring AI-driven workflows are reliable and continuously improving. On the agentic 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 multiple industry verticals. I also 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, Cohen's d effect size, and CI/CD integration with automated safety gates. These are not prototype experiments — they are production systems with statistical rigor baked in. ## From Research to Platform Leadership The thread connecting my ML research, my founder work, and my platform PM career at Intuit is the same: building the infrastructure that lets developers and AI systems move faster without sacrificing reliability. That is precisely what SoFi's AI SDLC platform requires — and why this role resonates. What specifically draws me to this role is the scope of the mandate: owning the multi-year product strategy for AI across the full product and software development lifecycle, from Spec-to-Code-to-Deploy workflows through CI/CD integration, observability, and governance. SoFi's position — with Technisys as the core banking substrate and Galileo as the B2B infrastructure layer — means that velocity improvements in the SDLC compound across consumer products, partner integrations, and internal platform teams simultaneously. The opportunity to define how AI agents integrate into specs, repositories, pipelines, and operational systems at that scale, inside a company with a genuine bank charter and real infrastructure depth, is not a feature PM job. It is a platform founding role, which is exactly how I operate. ## 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. Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma. - **Platform Infrastructure Scaling (Intuit):** Scaled throughput from 6K to 50K TPS via rSocket migration supporting approximately 1.5M concurrent connections with sub-25ms TP99 — the same order of infrastructure discipline SoFi's Real-Time Payments Platform requires. - **Java and Python SDK Starter Kits (Intuit):** Extended SDK scaffolding with build configurations (Gradle/Maven), testing frameworks, and CI/CD integration — enabling developers to go from zero to production-ready microservice in minutes. This is the developer experience layer that an AI SDLC platform must get right. - **MSaaS Drift Detection Program (Intuit):** 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 — directly analogous to the governance and standards enforcement this role requires. - **RL Workbench and aeval (AI/ML Projects):** Built production evaluation infrastructure for AI models — statistical rigor, CI/CD integration, automated safety gates, and cross-framework benchmarking — establishing the evaluation framework discipline the JD explicitly calls for. - **OpenClaw Multi-Agent Orchestration (StreamIO AI):** Designed and implemented a multi-agent gateway with subagent delegation, profile management, and session switching — hands-on experience with the agentic patterns (tool use, planning/execution loops, structured prompting) the JD identifies as required. - **Enterprise-Wide Service Language Assessment (Intuit):** Conducted analysis across 9 languages (Java, Python, Kotlin, Go, TypeScript, Scala, PHP, C++, Groovy), synthesizing usage data and developer feedback into strategic investment recommendations presented to the CTO — the kind of cross-organizational platform governance this role demands. ## Closing SoFi's mission — changing the way people think about and interact with personal finance — requires a development organization that can move at the speed of the market it is disrupting. An AI-powered SDLC platform is not a productivity tool; it is a strategic capability that determines whether SoFi can out-build incumbents across lending, banking, investing, and infrastructure simultaneously. I have spent 12 years building the technical depth, platform PM judgment, and AI systems experience to own that platform from day one. I would welcome the opportunity to discuss how my background maps to SoFi's specific ambitions here. Respectfully, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info