Notebook

What I tried, broke, and kept.

Technical essays and field notes from the messy middle: the question, the method, the evidence, and the thing I would do differently next time.

Index

ExperimentsResearchArgumentsLessons

Pinned note

A good first read from the notebook.

Pinned

Research Report12 min / December 2025

SEAS Search: building a knowledge graph-based course QA system

How we used knowledge graphs and LLMs to answer course-planning questions without letting the model invent prerequisites.

QuestionMethodEvidence

SEAS Search: building a knowledge graph-based course QA system

Inside

Question, evidence, analysis

More notes

Older pieces, sketches, and experiments I still stand behind.

9 pieces

Trying to Break the LLM Safety Barrier cover image

Experiment Log8 min

February 2026

Trying to Break the LLM Safety Barrier

My mission? See if I could get Google's Gemini or ChatGPT to give me genuinely dangerous advice. Not the 'don't try this at home' kind, but the 'this could literally harm my health' kind.

Prompt, method, result, takeawayAI Safety

Position Note7 min

January 2026

Women in STEM: Advocacy Through Action

The stat is real: women are underrepresented in STEM. I'm less interested in that conversation and more interested in what actually helped me stay.

Claim, reasoning, implicationsWomen in STEM

Experiment Log14 min

November 2025

2D to 3D camera calibration

How do you translate a GPS coordinate into a pixel on a webcam image? We explored camera calibration uncertainty using leave-one-out and Monte Carlo analysis on two real-world webcams. One urban, one on a remote harbor island.

Prompt, method, result, takeawayComputer Vision

Position Note6 min

June 2025

Bridging the Digital Divide with AI

Growing up in Nepal, I didn't have reliable internet until I was a teenager. Here's what I've actually seen AI do (and not do) for communities like the one I grew up in.

Claim, reasoning, implicationsAI

Bridging the Digital Divide with AI cover image

Experiment Log10 min

May 2025

Can a Generalist Model Learn to Specialize? Three Experiments with SAM

We took Meta's Segment Anything Model and tried to make it better at three specific tasks. One experiment made the model 36% worse. One improved it modestly. One was a breakthrough, and ended with a trip to the Smithsonian Zoo.

Prompt, method, result, takeawayComputer Vision

Research Report8 min

March 2025

The Intersection of UX and Machine Learning

How human-centered design principles can guide the development of more ethical and accessible AI systems.

Question, evidence, analysisUX Design

From Nepal to DC: A Journey in Tech cover image

Position Note5 min

October 2024

From Nepal to DC: A Journey in Tech

Reflections on growing up in Nepal, experiencing the digital divide firsthand, and pursuing computer science in the United States.

Claim, reasoning, implicationsPersonal Journey

Personal Note2 min

May 2018

happiness..

Taking time out for yourself, to be alone with yourself, to know your weaknesses and your strengths and to make yourself a better person is satisfying.

Reflection, context, lessonLife

Personal Note3 min

May 2017

Life

Life is not as easy as you dream it to be. Complications, vibes of giving up, sarcasm of people, demotivation from your loved ones, it comes eventually.

Reflection, context, lessonLife