applied AI studio — from mvps to apps to infra
even the most complex applied AI systems can be reduced to core components.
simpler to build. faster to value.
i design and build AI products, agentic systems, and trust layers like gatewaystack.
experience at aws and dell; educated at harvard and uchicago.
i help teams ship real AI systems fast using the openai apps sdk, mcp servers, and modern cloud infrastructure.
i build both the apps and the infrastructure behind modern agentic systems.
my app inner helps users collect dreams and memories, map emotions over time, and explore personal meaning through an agentic llm interface.
it’s available in the ios app store and inside chatgpt as an emotional and personal memory context layer for llms.
i built gatewaystack, secure user-scoped trust and governance for model and data access using the openai apps sdk and the model context protocol (mcp).
the gatewaystack modules form a composable architecture for user-scoped AI systems. each layer solves a foundational requirement of modern agentic applications:
identifiabl — bind every request to a real user and contexttransformabl — safety, normalization, classification, and preprocessingvalidatabl — permissions, policies, roles, and organizational ruleslimitabl — rate limits, quotas, budgets, spend controlsproxyabl — identity-aware routing across models and providersexplicabl — full audit trail, traces, and observabilityindividually useful but designed to interlock, gatewaystack defines the emerging trust and governance layer of the AI stack — the primitives every agent ecosystem will eventually rely on.