cover_letter / art_qqVK_teKQFs

Dear ElevenLabs Hiring Team, ElevenLabs is doing something genuinely rare: building AI voice infrastructure that doesn't just approximate human speech but redefines what's possible at the interaction layer between people and technology. That mission resonates directly with work I've been doing at Streamio AI, where I built a real-time HLS livestreaming pipeline and multimodal AI analysis system — and where the gap between a technically correct audio/video pipeline and one that actually serves users in production became very concrete, very fast. **Technical Foundation** My engineering background spans the full stack of AI-adjacent systems that matter for this role. At Streamio AI, I architected and shipped the OpenClaw multi-agent orchestration framework — a gateway protocol with subagent delegation, profile management, and session switching — coordinating AI agent workflows across multiple industry verticals. That system required the same kind of end-to-end ownership and integration depth that a Forward Deployed Engineer role demands: designing the architecture, writing the code, and then sitting with customers to understand where the seams showed. On the AI/ML side, I built a production RL post-training workbench that benchmarks GRPO, DPO, and 10 other algorithms (PPO, DAPO, REINFORCE, REINFORCE++, RLOO, SimPO, IPO, KTO, ORPO, SPPO) across TRL, VeRL, OpenRLHF, and NeMo RL — with live SSE metric streaming, GPU Docker passthrough, and cross-framework throughput/memory/convergence benchmarking. I also built aeval, a local-first model evaluation platform with FastAPI orchestration, TimescaleDB, Redis job queuing, and statistical rigor including bootstrap confidence intervals and Welch's t-tests. These aren't research prototypes — they're systems built to answer real questions about model behavior under production conditions. My NeurIPS 2014 paper on neural networks for protein secondary structure prediction — and the 2026 rewrite of that original 2004 C++ BPTT system into a PyTorch platform spanning 413 to 8B parameters — reflects a long arc of building AI systems from first principles, not just assembling frameworks. **Why This Role** The Forward Deployed Engineer role at ElevenLabs sits at exactly the intersection where I've operated most effectively: deep enough in the code to build real integrations, close enough to customers to understand what they actually need, and moving fast enough to matter. Having scaled Intuit's ICE platform to 675M+ engagements and 50K TPS while working directly with engineering teams across QuickBooks, TurboTax, and Mailchimp, I know what it takes to make complex infrastructure legible and usable for partners who have their own deadlines and constraints. **Role-Specific Connection** ElevenAgents, ElevenCreative, and ElevenAPI represent three distinct integration surfaces — each requiring a different technical conversation with a different kind of customer. What excites me about this role is the combination of hands-on execution and customer proximity: designing architectures with a client's engineering team in the morning and writing the integration code in the afternoon. ElevenLabs' stated emphasis on high-velocity, lean autonomous teams with minimal bureaucracy matches how I've operated as a founder and as a staff PM — I'm most effective when I own the outcome end-to-end rather than coordinating handoffs. **Selected Prior Experience** - Implemented OpenClaw multi-agent orchestration framework (gateway protocol, subagent delegation, profile management, session switching), coordinating AI agent workflows across real estate, insurance, health/dental, and financial-markets industries — directly applicable to ElevenAgents integration work. - Shipped Vantage on iOS and macOS with a FastAPI backend supporting 40+ agent tools, streaming agent chat via SSE, enforced mutation approvals, and idempotent endpoints — demonstrating production-grade API design and reliability engineering. - Architected a RAG retrieval pipeline (ChromaDB) with multi-provider LLM orchestration (Claude, GPT-4, Gemini), fallback routing, structured-output validation, and token-budget optimization for Fintellect AI — the same class of integration complexity ElevenAPI customers face. - Delivered ICE Self-Service platform (DevPortal, GitOps config, ICE Playground) at Intuit, reducing developer onboarding from 2–3 weeks to minutes in pre-prod and under 24 hours for production — a direct analogue to helping ElevenLabs' strategic partners go from evaluation to production deployment. - Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23; scaled throughput from 6K to 50K TPS via rSocket migration supporting ~1.5M concurrent connections with sub-25ms TP99 — evidence of operating at enterprise scale with real reliability requirements. - Built aeval with CI/CD integration, regression detection, and automated safety gates — the kind of evaluation infrastructure that strategic customers need before they can trust a voice AI system in production. - Led Mailchimp GCP-to-AWS migration for MSaaS at Intuit, delivering Golang template, MySQL persistence integration, and updated DevPortal documentation to meet production deadline — representative of the cross-functional, deadline-driven execution this role requires. **Closing** ElevenLabs is building the audio layer of the AI stack — the part that determines whether an interaction feels real or feels like a demo. That's consequential work, and the Forward Deployed Engineer role is where that work meets the customers who will actually deploy it at scale. I'd welcome the opportunity to discuss how my background in AI systems, developer platform infrastructure, and hands-on customer-facing technical work maps to what your most strategic partners need right now. Sincerely, **O. Felix Amoruwa** famoruwa@berkeley.edu | 909-731-9011 | felixamoruwa.info