# About

> I build production AI systems and study how autonomous agents can reason, remember, use tools, and operate reliably over long horizons.

Type: page
Canonical: https://nithinraphael.com/about
Status: published

---

**Location:** New York
**Focus:** Agentic AI, Memory systems, Retrieval & RAG, Distributed systems, Long-horizon autonomy
**Skills:** React & Next.js, TypeScript & Node.js, Go, Python, PyTorch & Transformers, LangGraph, PostgreSQL & Redis, Docker & Kubernetes, AWS & GCP

With over eight years of experience building in early-stage startups and shipping more than 35 products, my work has evolved from full-stack web architectures to highly concurrent distributed systems, and now to the frontier of agentic artificial intelligence.

My approach to engineering is fundamentally deeply structural. Whether designing sub-100ms real-time messaging pipelines in Go or orchestrating complex LangGraph environments backed by durable execution contexts like DBOS, I believe in building systems that fail gracefully, execute deterministically, and scale without friction.

Currently, my research focuses on the intersection of memory and autonomy. Vector databases are not enough for true reasoning; I am actively exploring hierarchical episodic-semantic dual-write architectures that allow AI agents to maintain consistent state across thousands of recursive execution steps.