Open to senior & AI roles

I build systems that stay fast and secure when it matters.

Senior full stack engineer, 10+ years across search, payments, access control and real-time features. Ask the terminal anything: it answers from my actual documents, with sources.

See how the terminal works >Get in touch >

ask.shgrounded in my docs

> ask"Biggest performance win?"

Rebuilt user search on OpenSearch vector embeddings. Average response: 5s → <200ms.

try

Questions are logged without personal identifiers so I can see what people ask.

Previously building at

How the terminal works

A small RAG system, built like production.

Pinecone retrieval through LangChain, answered by Claude from a Nuxt server route. Scoped to questions about me, defended against prompt injection, rate-limited, and every answer cites the document it came from.

rag.pipeline.ts

pipeline({
  sources: ["resume", "projects", "build-writeup"],
  chunk:   "by-section",
  embed:   PineconeEmbeddings("multilingual-e5-large"),
  store:   PineconeStore,
  guards:  ["scope", "injection", "rate-limit"],
  answer:  { model: "claude", cite: true },
})

Open source

Small tools, published on npm.

tailwind-sidebar-expandedExpandable/collapsible sidebar variant for Tailwind.ngx-countdownAngular directive for configurable countdown timers.no-logStrips console logging from production builds.