Hi, I'm

Hardik Chaudhary

AI graduate building skills in agentic AI and AI-driven workflows.

I build AI-assisted products like InThink, and I'm working toward a remote role in agentic AI or AI integration once I graduate in August 2027.

View Projects Download CV

Good AI workflows still start with understanding people, not tools — so before I ask what something should do, I ask why anyone would trust it.

Featured

InThink

InThink started as a way to answer a question I kept asking myself: why do good ideas from books and conversations disappear again within a week? It's a daily-concept learning app spanning psychology, philosophy, sociology and relationships, built around spaced re-encounter and Socratic "teach it back" prompts so concepts actually stick instead of being read once and forgotten. Alongside it sits Think, an AI thinking partner, and a private notebook — all built on a content model that leads with feeling before explanation, because that's usually the order people actually process new ideas in. I designed, wrote, and built it solo, mixing AI tools deliberately rather than leaning on one for everything.

Visit InThink ↗

Projects

InThink, plus a couple of smaller builds alongside it.

Live · In Testing

InThink

A daily-concept learning app covering psychology, philosophy, sociology and relationships, with a private notebook and an AI thinking partner called Think — see the story above.

Demonstrates: product design + AI-assisted solo build

Live ↗
Live

This Site

This portfolio itself — designed and built hands-on with Claude Code. A working agentic-AI project rather than a resume line, iterated on directly with an AI coding partner.

Demonstrates: AI-assisted engineering workflow

Live ↗
Published

Notes MCP Server

A Python MCP server that lets Claude read, search, create, append to and safely delete markdown notes. Claude chooses the arguments for each tool call, so the server enforces its own limits: path-traversal protection, soft-delete to a recoverable .trash folder, and 12 automated pytest tests.

Demonstrates: MCP tool design, input validation, automated testing

GitHub ↗ Watch demo ↗
Published

File Clutter Archiver

A Python command-line tool where fixed rules or Claude propose which files are clutter, and nothing moves until you approve each one. Claude can only propose, files are archived and never deleted, and every decision is logged. 21 automated tests run without any API calls.

Demonstrates: AI agent tool use, human-in-the-loop design, safety guardrails, testing

GitHub ↗ Watch demo ↗

Next Build

Something new is in early planning — check back soon.

About & Approach

I'm a BSc Computer Science (Software Engineering) undergraduate at the University of Hertfordshire, and outside of coursework I've been building InThink solo — deliberately mixing AI tools rather than leaning on one for everything: one for research and synthesis, another for interface generation, another for strategy. Figuring out which tool earns its place in that stack is half the actual work.

What keeps pulling me back to this space is a genuine interest in human behaviour and cognitive bias — confirmation bias, self-serving bias, cognitive dissonance. I try to steel-man my own product and content decisions before I trust them, and that habit is really where my interest in human-centred AI comes from, not something I picked up because it sounds good on a CV.

Right now I'm going deep on the broader agentic AI landscape — Claude Code, Claude Cowork, autonomous and self-hosted agents, no-code AI builders — working toward a remote role in AI orchestration or AI operations once I graduate in 2027.

Timeline

How this has actually unfolded, in order.

  1. Sep 2024 – May 2027 (expected)

    University of Hertfordshire

    BSc Computer Science, Software Engineering — coursework spanning software engineering, systems, and AI-adjacent modules.

  2. Jun 2026 – ongoing

    InThink

    Founded and solo-built a daily-concept learning app and AI thinking partner, from product design through to shipping.

  3. Aug 2026 – ongoing

    Learning Agentic AI Tooling

    Going deep on Claude Code and agentic AI workflows — autonomous agents, orchestration, and AI-driven automation.

  4. May 2027

    Available for Full-Time Remote Roles

    Open to remote positions in AI orchestration or AI operations after graduating.

Currently Learning

What I'm going deeper on next, building on the timeline above.

Want to work together? Let's talk.

Open to remote roles in agentic AI and AI integration from May 2027.

Email hardikchaudhary267@gmail.com LinkedIn in/hardikchaudhary1111