cover_letter / art_VNlsAy9iNDg

Dear Databricks Hiring Team, Databricks sits at the intersection of where enterprise data infrastructure and production AI converge — and the Agentic Enterprise Applications Platform represents exactly the kind of foundational, developer-facing work that determines whether AI actually reaches production or stalls in prototype. My interest in this role is grounded in a specific experience: building the ICE Self-Service platform at Intuit, where I reduced developer onboarding from 2–3 weeks to minutes by designing a self-service DevPortal, GitOps configuration layer, and playground environment — the same architectural instincts this role demands at the agent runtime and MCP connector layer. ## Technical and AI Foundation My technical credibility in agentic AI is hands-on and recent. At Streamio AI, 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. This is not a prototype; it runs in a production Electron + React + TypeScript application with Claude MCP SDK integration, real-time screen capture analysis, and a cross-platform deployment pipeline including macOS code signing and notarization. I understand what an agent runtime is because I built one. On the evaluation side, I built aeval — a local-first AI model evaluation platform with five core eval types (factuality, reasoning, instruction-following, safety, code generation), adversarial safety testing with refusal detection, and statistical rigor including bootstrap confidence intervals, Welch's t-test, Cohen's d effect size, and saturation detection. The platform integrates CI/CD regression detection and automated safety gates — a direct analog to the mandatory evaluation gates in the promotion pipeline this role owns. The stack: FastAPI orchestrator, TimescaleDB, Redis job queue, Next.js dashboard, Ollama. My RL post-training workbench goes further: a 3-phase platform covering Reward Lab (A/B testing reward functions across GSM8K, MATH, HumanEval, UltraFeedback), a TRL-powered training Playground with live 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 with standardized throughput, memory, and convergence benchmarking. This maps directly to Databricks' Mosaic AI model training and evaluation infrastructure. At the research layer, my NeurIPS 2014 paper on neural networks for protein secondary structure prediction — and the 2026 rewrite of that system in PyTorch spanning 413 parameters to 8B (a 19-million-fold scale increase) with MLflow experiment tracking, Optuna HPO, and FastAPI serving — establishes that my engagement with ML is not surface-level. ## Why This Role The through-line of my career is building platforms that other teams build on: the ICE framework at Intuit (675M+ engagements, scaled from 6K to 50K TPS via rSocket migration supporting ~1.5M concurrent connections), the Search Service and SPL/SPL2 ecosystem at Splunk, and now agentic AI infrastructure at Streamio and Fintellect. The Databricks Agentic Platform role is the natural next chapter — owning the foundational layer that domain teams across GTM, Finance, HR, and Legal depend on, with the technical depth to engage on agent runtime architecture, context graph design, and evaluation pipeline construction. What specifically draws me to this role is the three-layer intelligence architecture — knowledge graph, context graph, and temporal memory — and the requirement for unified retrieval across vector, structured, and graph sources with source traceability. At Fintellect AI, I 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. That experience gives me direct intuition for the trade-offs in the intelligence layer this role defines. ## 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. Scaled to 675M+ engagements in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma. - **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. Directly analogous to the connector SDK and agent templates this role ships. - **OpenClaw Multi-Agent Orchestration (Streamio AI):** Built gateway protocol, subagent delegation, profile management, and session switching for coordinated AI agent workflows — production implementation of the managed agent runtime concept central to this role. - **aeval Evaluation Platform:** Built AI model evaluation with adversarial safety testing, automated safety gates, CI/CD regression detection, and statistical rigor (bootstrap CIs, Cohen's d). Direct precedent for the AI-judge evaluation pipeline and mandatory promotion gates this role owns. - **RAG Pipeline (Fintellect AI):** Architected retrieval pipeline with ChromaDB, multi-provider LLM orchestration with fallback routing, structured output validation, and token budget optimization — relevant to the intelligence layer and context retrieval architecture. - **Search Orchestration (Splunk):** Owned Search Service (Go microservices), Search Catalog (PostgreSQL metadata service), and SPL/SPL2. Delivered Scheduler Service end-to-end in ~4 months; led query performance optimization achieving up to 10x improvements in Splunk Cloud Services search. - **MSaaS Drift Detection (Intuit):** Wrote Java JAR library to scan Git repos for configuration drift and built remediation roadmap using OpenRewrite — the same governance and quality-gate instincts required for the promotion pipeline from prototype to production. ## Closing Databricks' mission — unifying data, analytics, and AI so that every organization can build on a single, governed intelligence platform — is the right problem to be working on. The Agentic Enterprise Applications Platform is the layer that determines whether that mission reaches the teams who need it most. I bring the platform product experience, the hands-on agentic AI implementation depth, and the cross-functional execution track record to own this roadmap and ship it. 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