Richard Rizk

Richard Rizk is Co-Founder and CTO of BLACKBOX AI — and the pioneer behind the only end-to-end encrypted inference technology available today. He built Encrypted Model: client-sealed AI inference where prompts are encrypted on the user’s device and decrypted only inside an attested GPU enclave. The proxy, the host, and the operator never hold the key. AES-256-GCM. ECDH P-384. Hardware attestation before secrets move. True end-to-end encryption for AI inference and agentic workloads — the Signal standard, applied to the model hop. Where others stop at TLS or trust the gateway, Richard’s stack answers a harder question: can anything except you and the attested enclave read the prompt at all? At WeAreDevelopers World Congress North America, he joins leaders from Sentry, Render, and Cognition on the mainstage to discuss what it takes to run AI-written software in production — at speed, and without surrendering the prompts that built it.