███ ██ █████ █████
░░░ ███ ░░███ ░░███
████ ░░░ █████████████ █████ █████ ██████ ███████ ██████ ████████ ███████
░░███ ░░███░░███░░███ ░░███ ░░███ ███░░███ ███░░███ ░░░░░███ ░░███░░███ ░░░███░
░███ ░███ ░███ ░███ ░███ ░███ ░███████ ░███ ░███ ███████ ░███ ░███ ░███
░███ ░███ ░███ ░███ ░░███ ███ ░███░░░ ░███ ░███ ███░░███ ░███ ░███ ░███ ███
█████ █████░███ █████ ░░█████ ░░██████ ░░████████░░████████ ████ █████ ░░█████
░░░░░ ░░░░░ ░░░ ░░░░░ ░░░░░ ░░░░░░ ░░░░░░░░ ░░░░░░░░ ░░░░ ░░░░░ ░░░░░
SEE.mp4
Building intelligent things and occasionally breaking them.
I'm Vedant, an AI/ML developer focused on building systems that go beyond simply calling an LLM and printing its response.
I'm currently diving deep into:
🤖 Generative AI & LLM Applications 🧩 LangChain & LangGraph 🔀 Multi-Agent Systems 🧠 RAG & Long-Term Memory 🗣️ NLP & Fine-Tuning ⚙️ AI Automation & Tool Calling 🚀 Turning ideas into actual working projects
My approach is simple:
Understand the problem → design the system → build it → break it → understand why → build it better.
And yes...
Sometimes the last step is just
works on my machine™.
A darker interface for the late-night engineering sessions where one bug somehow turns into a three-hour debugging adventure.
I'm also interested in how the visual layer of an AI application can make complex systems feel easier and more intuitive to use.
Because let's be honest:
Some of the best debugging happens after midnight. 🌙
Exploring better ways for humans to interact with AI systems through commands, inputs and responsive interfaces.
Because a powerful AI system is only useful if interacting with it doesn't feel painful.
The interface should make it clear what the user can ask, what the system understands, and what happens after an action is triggered.
Not everything should silently happen in the background.
I'm interested in systems that can surface important events, anomalies and decisions instead of hiding everything behind a terminal.
For intelligent systems, observability matters.
It's useful to know:
- what happened
- why it happened
- what the system decided
- what failed
- and what should happen next
If the system makes a decision, I want to be able to understand it.
These views represent another part of what I'm learning:
AI systems shouldn't just be intelligent under the hood — they should also be understandable from the outside.
Good interfaces make it easier to see:
- what the system is doing
- what it knows
- what it decided
- what happened during a session
- and where something went wrong
I'm particularly interested in connecting the backend reasoning layer with the interface layer, so that the user isn't completely blind to what the system is doing.
Coding is a big part of what I do.
But not the only part.
I like exploring new places, collecting experiences, experimenting with ideas and occasionally disappearing into a completely unrelated rabbit hole for several hours.
Because sometimes the best ideas don't come from another tutorial.
They come from doing something different.
