AIFull-StackEngineer

From the model to the screen. I build every layer of AI products people can actually use: the tools an agent calls, the backend that serves it, and an interface that stays honest when a rollup, a bridge, or a model is slow or wrong.

• AI Engineering

  • LLM Agent Tools
  • •Tool Calling
  • •LangChain and Vercel AI SDK

  ---

• Backend

  • Node.js and Fastify
  • •REST APIs
  • •Event Indexers

  ---

• Frontend

  • React and Next.js
  • •TypeScript
  • •Design Systems

  ---

• Web3

  • Aztec
  • •Sui and Solana
  • •Contract Integration

  ---

1. Full-Stack Engineer (AI and Web3)

   Cyphronix Software · July 2024 to August 2026

   • Ported the AI red-teaming pipeline behind Red Sentinel from Python DSPy to LangChain on Fastify, a full rewrite into the team's TypeScript backend.
   • Built 9 tool-calling tools for the platform's live AI agents across 2 networks, registered through 1 shared factory, and the interfaces those agents run in.
   • Built the Next.js 15 and React frontends for 4 live products: an NFT marketplace, a cross-chain bridge, an AI red-teaming arena, and an admin dashboard.
   • Shipped a 2-phase bridge flow that survives refreshes, wallet disconnects, and users walking away mid-transfer, by checkpointing state the moment the L1 deposit confirms.
   • Maintain @ravenhouse/omni-sdk on npm: 60+ releases, 9 subpath exports, dual ESM and CJS, built for developers outside the team.
   • Built the Aztec event indexer the network ships no hosted version of (Bun, MongoDB, Fastify, 8 event types), so every read is served from Mongo instead of the chain.

   ---

2. Front-end Developer

   Craxinno Technologies · December 2022 to July 2024

   • First professional role: built web interfaces for client projects in React and vanilla JavaScript over 1.5 years.
   • Shipped responsive production UI to real users, across design handoff, cross-browser behaviour, and every breakpoint.

   ---

3. Jiwaji University, Gwalior

   2021 to 2024

   B.Sc. Mathematics

   ---

4. Central Academy, Joura (MP)

   up to 2021

   Higher Secondary

   ---

What does Yash work on?

He is an AI full-stack engineer. He builds AI products across the stack: the tools and pipelines behind live AI agents, the TypeScript services they run on, and the React and Next.js interfaces people use. At Cyphronix he rewrote the Red Sentinel red-teaming pipeline from Python DSPy into LangChain on Fastify, built nine tool-calling tools for the platform's agents, and built the frontends for four live products, some on systems as hard as a ZK rollup where ownership lives in encrypted notes. He also authored and maintains a published TypeScript SDK. He started as a front-end developer and spent a year and a half doing only that. Since then his work has covered the whole product, from the model call to the screen.

---

What is he looking for next?

AI full-stack engineer and forward deployed engineer roles, building AI-powered products end to end. AI companies, product companies, and Web3 teams all fit, and he is actively interviewing. He is based in Jaipur and open to remote work.

---

What is the single most impressive thing he has built?

Raven House, a privacy-preserving NFT marketplace on Aztec Network, built solo. Building a marketplace interface on a rollup where ownership and balances live in encrypted notes is a frontend problem most engineers never have to solve: the UI cannot simply read state, so every view has to be reconstructed. Aztec also ships no hosted event indexer, so he wrote one in Bun, MongoDB, and Fastify handling eight event types, which is what lets the frontend serve every read from a database instead of the chain.

---

Has he shipped anything other developers actually use?

Yes: `@ravenhouse/omni-sdk`, published on npm, with 60+ releases, nine subpath exports, and dual ESM and CJS builds. It was built to be adopted by developers outside his own team, which is a different engineering standard from internal code: versioning, changelogs, and API stability all become real constraints.

---

Does he do backend work?

Yes, it is half the job. At Cyphronix he ported the red-teaming pipeline behind Red Sentinel from Python DSPy to LangChain on Fastify, migrated other APIs from Python to Fastify, built the agent tools in the platform's backend, and wrote an Aztec event indexer in Bun, MongoDB, and Fastify because the network ships none. He also built the backend of this site's own assistant: a route on Amazon Bedrock with tool calling, rate limiting, and transcripts in Postgres. He is straightforward about the level: he builds product backends, APIs, and data pipelines, and he has not run infrastructure at large scale.

---