AI Researcher

Speaker

Vibe Coder

Vanchhit
Khare

People call me Van.

Professional
Introspector

Always eager to learn

Introspector

Two shifts. One obsession.

9 to 5, I’m a solutions engineer.

5 to 9, I build AI systems nobody asked for but everyone ends up needing.

Weekends, I argue with my AI agents. They usually win.

Drunk or sober, your best friend and your honest critic.

1

The
Numbers

The Numbers

All real. I counted.

16xSalesforce Certified
10Papers written
60+Papers reviewed
MSCal State Fullerton

Papers shown at conferences in

USA, Canada, Egypt, Romania, New Zealand.

Reviewed for

IEEE, Springer, Elsevier, MDPI, IGI Global, OpenReview.

2

My next judging

Carnegie Mellon University Pittsburgh

NOVA

Judging

They invited me. On purpose.

NOVA

CMU hackathon. Invited by ScottyLabs, Pittsburgh.

Cal Hacks

UC Berkeley, June 20-21 2026.

3

Speaking

Mentorship

Mentorship

Pizza apps don’t make it.

AI4ALL Ignite

Guided students, several left LinkedIn recommendations.

Backed by Melinda French Gates, NVIDIA, Accenture

TechBridge.org

Backed by Chevron, Moore Foundation, NSF

AI Ventures by Technovation

Mentoring two teams: road-safety in Asia and an AI VR birth-prep tool in Nigeria.

Backed by UNICEF

All Star Code

Mentor and instructor, 2026.

5

Published
Writing

Now

Now

Five threads. Zero chill.

AI Second Brain

A memory my AI agents can read from and add notes back to.

My robots’ little notebook.

Context Engineering

I shrink an agent’s memory at 75% full so it never runs out of room.

I tidy the robot’s messy desk.

Loop Engineering

Agents that keep going on long jobs without losing track.

Robots that never give up.

Model Collapse

How AI gets worse when trained on AI-made data, and how to stop it.

Stops AI eating its own tail.

Token Reduction

I cut how much an agent reads per task so it costs less.

Robots read less, cost less.

7

Let’s talk.

The end. No sequel yet.

Millennial by birth. Gen Z by energy.

Cover