cover_letter / art_XrlIR3Xb1GQ

Dear Okta Identity Team Hiring Manager, Okta sits at the precise inflection point where identity infrastructure becomes the trust layer for an AI-native world — a mission that carries genuine stakes as autonomous agents proliferate and session-based attacks grow more sophisticated. My path from hand-coding backpropagation through time in C++ at UC Berkeley in 2004 to building multi-agent orchestration frameworks and RL post-training workbenches in 2026 has consistently placed me at the intersection of deep technical systems and product strategy — exactly the intersection this role demands. ## Technical Foundation My AI and developer platform credentials are grounded in production systems, not prototypes. At Intuit, I owned the developer frameworks and platform infrastructure that scaled to 675M+ engagements in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma — including a rSocket migration that pushed throughput from 6K to 50K TPS while supporting approximately 1.5M concurrent connections at sub-25ms TP99. I extended Java and Python SDK Starter Kits with scaffolding, CI/CD integration, and testing frameworks, and delivered the ICE Self-Service platform that compressed developer onboarding from two to three weeks down to minutes in pre-production. That experience — owning a platform that thousands of internal developers depend on, with real revenue consequences for reliability — is directly analogous to the "Okta on Okta" Customer Zero mandate described in this role. On the agentic AI 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 industry verticals. The gateway pattern I implemented maps directly to the identity and security gateway architecture Okta is defining for O4AA and A4AA. I also architected a RAG retrieval pipeline with multi-provider LLM orchestration (Claude, GPT-4, Gemini) with fallback routing, structured output validation, and token budget optimization — giving me hands-on fluency with the authentication and authorization patterns that AI agent workflows require. My aeval platform — a local-first model evaluation system built on FastAPI, TimescaleDB, Redis, and Ollama — demonstrates the kind of rigorous, metrics-driven thinking this role requires: bootstrap confidence intervals, Welch's t-test, Cohen's d effect size, automated safety gates, and CI/CD regression detection. Defining and monitoring KPI-based reporting for a security portfolio is a natural extension of that discipline. My NeurIPS 2014 publication on neural networks for protein secondary structure prediction, and the subsequent 2026 rewrite of that system spanning 413 parameters to 8B (a 19-million-fold scale increase), establish that my technical engagement with AI is longitudinal and substantive — not surface-level. ## Why This Role The convergence of agentic identity governance, hardware-bound session security, and continuous threat posture management represents the most consequential product surface in enterprise security right now. Okta's position — building the neutral infrastructure layer that governs both human and non-human identities — is the right architectural bet, and the "drinks our own champagne" Alpha-to-EA lifecycle is exactly the kind of Customer Zero rigor that separates durable security products from ones that fail in production. I want to be the person who owns that internal deployment loop and translates it into roadmap signal. Specifically, the Model Context Protocol (MCP) dimension of this role is one I have direct hands-on experience with: StreamIO's production deployment uses the Claude MCP SDK to expose screen capture tools to AI coding assistants, and I have implemented MCP server integrations in a shipping desktop application. Understanding MCP not just conceptually but as a practitioner gives me an immediate advantage in defining the token exchange patterns and agent-to-data security guardrails Okta is architecting. The anti-session hijacking work — Device Bound SSO, Chrome DBSC, cryptographic hardware-binding via TPM and Secure Enclave — is an area where my background in platform infrastructure, developer-facing SDKs, and security-adjacent systems (I led Splunk Logging-as-a-Service at Kaiser Permanente handling 1.7 TB daily volume across 200+ enterprise customers) gives me the technical vocabulary to engage engineering teams as a credible peer, not just a requirements translator. ## Selected Relevant Experience - **Intuit — ICE Self-Service Platform:** 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 owning internal rollout of Alpha features through Early Availability. - **Intuit — Platform Scale:** Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23; drove rSocket migration supporting ~1.5M concurrent connections at sub-25ms TP99 — demonstrating the infrastructure judgment required for enterprise-grade identity systems. - **StreamIO AI — OpenClaw Multi-Agent Orchestration:** Built gateway protocol, subagent delegation, profile management, and session switching for coordinated AI agent workflows — directly applicable to defining agentic governance guardrails and downscoped runtime permissions. - **StreamIO AI — MCP SDK Integration:** Shipped production MCP server exposing screen capture tools to AI coding assistants, with Kinde OAuth 2.0 authentication and Electron SafeStorage for secure credential management — hands-on experience with the exact protocols Okta is governing. - **Fintellect AI — Multi-Provider LLM Orchestration:** Architected RAG pipeline with ChromaDB, multi-provider LLM fallback routing (Claude, GPT-4, Gemini), structured output validation, and token budget optimization — practical fluency in how AI agents handle authentication and authorization. - **Splunk — Search Orchestration PM:** Owned Search Service (Go microservices), Search Catalog (PostgreSQL metadata), and SPL/SPL2; delivered Scheduler Service end-to-end in ~4 months; led query performance optimization achieving up to 10x improvements — demonstrated ability to ship complex technical platform features on compressed timelines. - **Kaiser Permanente — Splunk Logging-as-a-Service:** Led enterprise rollout handling 1.7 TB daily volume across 200+ internal customers, and built caching capability using Redis and XC10 for scalability and fault tolerance — enterprise security infrastructure at scale. - **aeval — AI Model Evaluation Platform:** Built adversarial safety testing with refusal detection, data contamination detection via SHA-256 hashing, and automated safety gates with CI/CD regression detection — the metrics and safety rigor mindset directly applicable to Identity Security Posture Management. ## Closing Okta's mission — securing every identity connection and building the foundational trust layer for the modern enterprise — is not abstract to me. The systems I have built, from multi-agent gateways to developer platform infrastructure to model evaluation pipelines, all depend on identity and authorization being correct. Getting it wrong has real consequences. I operate with that weight, and I bring the technical depth, platform PM track record, and strategic judgment to help Okta define what agentic identity governance looks like before the rest of the industry catches up. I would welcome the opportunity to discuss how my background maps to the specific challenges your team is navigating. Sincerely, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info