Building practical AI systems

I'm Chirag Borse, an ai systems builder from Mumbai, India. I build agent tooling, terminal-native workflows, Graph-RAG memory systems, and open-source experiments around the modern AI development stack.

Builder Profile

AI infrastructure over wrappers

My current focus is building systems that make AI agents more useful, inspectable, and safe to run in real developer workflows. That includes orchestration runtimes, MCP ecosystem tooling, workflow dashboards, and memory architectures.

I like terminal-first experiences because they are fast, scriptable, and close to how developers already work. Most of my projects are built in public as practical experiments rather than polished marketing demos.

This portfolio is a snapshot of that direction: AI assistants, Graph-RAG memory, local-first observability, MCP security, and creator workflow tools.

Platform Features

Built for Modern Development

AI Agent Systems

Building assistants, orchestration patterns, and control interfaces for useful agent workflows.

Memory & Retrieval

Exploring Graph-RAG, persistent memory, entity relationships, and hybrid vector search.

MCP Tooling

Creating developer tools around MCP configs, agent credentials, and safer local setups.

Terminal Workflows

Designing command-line experiences that are inspectable, automatable, and fast to operate.

Local-first Systems

Using simple APIs, SQLite, Docker, and self-hosted stacks where transparency matters.

Building in Public

Shipping projects on GitHub, learning through prototypes, and turning experiments into tools.