cover_letter / art__I2Nz7EWLeM

Dear Hinge Health Hiring Team, Hinge Health is doing something genuinely difficult: making high-quality musculoskeletal care accessible to people who would otherwise wait months for a physical therapy appointment or skip treatment entirely. The intersection of clinical rigor and consumer product design is where the hardest and most meaningful product problems live. My path from hand-coding backpropagation in C++ at UC Berkeley to building multi-agent AI orchestration frameworks and scaling a developer platform to 675M+ annual engagements at Intuit has been oriented toward exactly this kind of challenge — building systems that are technically sophisticated but experienced by users as effortless and personal. **Technical and AI Foundation** My AI work is hands-on and spans the full stack. In 2025–2026, I built *aeval*, a local-first model evaluation platform with five core eval types (factuality, reasoning, instruction-following, safety, and code generation), adversarial safety testing with refusal detection, and statistical rigor built in — bootstrap confidence intervals, Welch's t-test, Cohen's d effect size, and automated safety gates integrated into CI/CD. That work reflects how I think about AI in products: not as a feature to ship, but as a system to measure, iterate on, and trust. For Fintellect AI, I 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 — then embedded domain-specific conversational agents into the mobile app, each scoped to a distinct financial focal point. This is the pattern Hinge Health needs: AI that is personalized, context-aware, and clinically bounded, not a generic chatbot bolted onto a product. For StreamIO, I implemented the OpenClaw multi-agent orchestration framework with a gateway protocol, subagent delegation, and session management — enabling coordinated AI workflows across distinct industry verticals. The architecture required the same thinking that personalization at scale demands: routing the right context to the right agent at the right moment, without exposing the complexity to the user. 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), establish that my engagement with ML is not surface-level. I understand model behavior, training dynamics, and evaluation — which makes me a more credible partner to engineering when making AI product decisions. **Why This Role** Hinge Health's consumer product challenge is one I find compelling precisely because it sits at the intersection of behavioral engagement, clinical outcomes, and AI personalization — three domains where I have direct experience. The opportunity to own a consumer-facing surface that drives retention and clinical results, while integrating generative AI into the treatment experience, is the kind of 0-to-1 and 0-to-scale problem I have been building toward. **Role-Specific Connection** The JD's emphasis on AI-native product development and personalization at scale maps directly to work I have already shipped. Building recommendation and routing logic that adapts to user context in real time — whether that is a retail investor's risk profile in Fintellect or a developer's service configuration in Intuit's ICE platform — requires the same muscle: forming a clear hypothesis, instrumenting the right metrics, and adjusting mid-flight. At Hinge Health, applying that discipline to member engagement and clinical adherence is a problem I am ready to own end-to-end. **Selected Prior Experience** - **Fintellect AI:** Architected RAG retrieval pipeline with ChromaDB, multi-provider LLM orchestration (Claude, GPT-4, Gemini) with fallback routing, and domain-specific conversational agents embedded in mobile and web — delivering context-aware, personalized advisory interactions at the user level. - **StreamIO AI:** Implemented OpenClaw multi-agent orchestration framework with gateway protocol and subagent delegation; built end-to-end auth and payments pipeline (Kinde OAuth 2.0, Stripe tiered subscriptions); integrated ElevenLabs TTS/STT with 6 auto-classified voice profiles — demonstrating 0-to-1 consumer product execution across the full stack. - **Intuit — ICE Platform:** Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23 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 — demonstrating metric accountability and platform thinking at consumer scale. - **Intuit — Developer Onboarding:** Delivered ICE Self-Service platform (DevPortal, GitOps config, 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 retention and activation problem with the same structure as member onboarding in a consumer health product. - **aeval:** Built model evaluation platform with adversarial safety testing, refusal detection, data contamination detection, and automated safety gates — establishing the evaluation infrastructure needed to deploy AI responsibly in a clinical context. - **Splunk — Search Orchestration:** Delivered Scheduler Service end-to-end in approximately four months; led query performance optimization achieving up to 10x improvements for a beta enterprise customer; designed RICE-based prioritization framework balancing multiple stakeholder requirements across three microservice backlogs. - **Kaiser Permanente:** Led enterprise rollout of Splunk Logging-as-a-Service handling 1.7 TB daily volume across 200+ internal customers; built Redis-based caching capability addressing scalability and fault tolerance — six years of healthcare-adjacent enterprise product experience with real operational stakes. **Closing** Hinge Health's mission — making world-class musculoskeletal care accessible and effective for people who need it — is one that compounds over time. Every member who completes their care plan, every engagement loop that keeps someone moving, every AI interaction that feels genuinely helpful rather than generic: these outcomes matter beyond the product metrics. I want to build the consumer experiences that make those outcomes more likely, at scale, with the rigor and craft the mission deserves. I would welcome the opportunity to discuss how my background maps to the specific challenges your consumer product team is navigating. Sincerely, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info