Hello, World.

Hey there! This is my first blog post, and honestly, I’m still figuring out what this space will become. But if you’re here, you’re probably interested in AI, open source, or how things break in production. Welcome.

Who’s Writing This?

I’m Gauresh Tambe, a 2025 Computer Science grad (AIML specialization) from Finolex Academy of Management and Technology. I graduated with a 9.32 CGPA and secured Rank 14 in Mumbai University—which sounds incredibly impressive until you realize I spent my entire final semester stressing over my final project and desperately trying to keep up with internship sessions.

Currently in my hometown, building AI tools and open-source libraries between job applications and LeetCode sessions.

What I’m Building (Right Now)

Docsy — Headless RAG for Documentation

If you are building your portfolio or a documentation site and want to add an AI chatbot—but don’t want to waste time building a production-grade RAG pipeline from scratch—Docsy is your solution.

Right now, it scrapes GitHub .md files (like your README.md). If you have built out a solid README and just want a quick, intelligent chatbot around your project without having to write a massive, dedicated documentation site, Docsy handles the heavy lifting for you.

What it does right now:

• Scrapes GitHub .md files directly.
• Ingests data seamlessly using the CLI.
• Supports both Naive and Advanced RAG patterns out of the box.

I am currently building out new, complex RAG architectures and adding a web scraping feature to expand its capabilities.

Check it out: docsy.live 

query.fit — AI-Native Database Assistant

I originally built query.fit because I wanted to be able to simply “talk” to my databases. It started strictly as a deep-dive learning project to wrap my head around LangGraph and agentic architecture, and it quickly evolved into a fully functional tool.

What it does: It acts as an AI-powered interface that translates natural language into executable SQL. You ask a question, and the underlying agent generates, validates, and executes the query to return your data (along with dynamic chart visualizations).

Key Features:

• Supports local SQLite files, plus live PostgreSQL, MySQL, and MariaDB connections.
• Orchestrates complex agentic workflows using LangGraph to ensure the generated SQL is actually valid before running.
• Handles secure connection routing with encrypted credentials and auto-expiring sessions.
• Wrapped in a clean, accessible interface built with Shadcn UI.

Tech Stack: Next.js, Supabase, LangChain, LangGraph, Shadcn UI

What You’ll Find Here

I’ll be writing about whatever’s on my mind:

• AI & Tech: RAG pipelines, LLM experiments, open-source projects
• Code & Learning: Building in public, debugging stories, lessons from shipping software
• Future Thoughts: Where AI is heading, tools I’m excited about
• Self-Growth: Balancing work and life, Calisthenics, learning in public
• Life Stuff: Gaming, random hobbies when I’m not staring at code

Short posts, long posts, whatever fits. No fixed schedule.

Why This Exists

Most dev blogs either disappear after three posts or turn into marketing disguised as tutorials.

I wanted a space to document what I’m building—the wins, the bugs, the stuff that breaks at 3 AM. No “10x developer” posturing. Just honest progress logs and things I wish I’d known earlier.

If that sounds interesting, stick around.

Let’s Connect

If you’re building something cool, experimenting with RAG, or just want to talk about AI, tech, codes, opensource let’s connect.

• GitHub: @GaureshArt  (star Docsy if you’re feeling generous)
• Twitter/X: @gaureshart  (mostly tech, occasional shitposts)
• Email: gauresh.art.ai@gmail.com (for serious stuff)
• LeetCode: G_Art  (if you want to judge my contest performance)

If you find a bug in Docsy, please open an issue. If you fix it yourself and send a PR, I’ll probably build a shrine in your honor.

Thanks for reading. See you in the next one. ✌️