Projects
Completed
- Write a children’s book using AI December 2024
One of the early use cases I thought I could monetize with AI was customized children’s books.
I envision a platform where individual users can enter 5–10 data points and be delivered a customized children’s book teaching a targeted and customized lesson to a child in need.
Developed a workstream and built one custom book before realizing that while the barrier to production has been effectively removed by AI, the barrier for market share remains. Learned how to sell books on Amazon, but sold a grand total of zero.
Dealt with the challenges of maintaining character consistency, and lucked my way into The Moonlight Keeper — something I genuinely enjoy.
Download the PDF and tell me what you think! I’ve been told that children find the design tone a bit overwhelming, an indicator that perhaps I lack other core skills necessary to be a children’s writer.
Let the end product represent itself!
The Moonlight Keeper — final PDF, free to view · under construction
- Write a book using AI 2025
I admit — I added in the ‘using AI’. My original vision was to write a book myself. But my lazy mind quickly spotted the potential of using AI to rapidly help me churn out a book. I wasn’t displeased with the final results myself, but this didn’t quite take the presses by storm.
Contents · under construction
- Submit an essay for the Berggruen Prize essay competition 2025
This was the first time in a while that I wrote with a consistent objective in mind for weeks on end. It was stressful, but I managed to finally produce something. Submitted in September 2025, I believe my essay to contain a strong vision for designing a symbiotic future with an alien race.
Titled ‘On designing first contact’ — “That this non-human intelligence is actually being created by human intelligence offers a unique opportunity for us to design our First Contact - a luxury rarely made available to civilizations.”
The essay, and a note on how it took shape across drafts · under construction
Ongoing
- The Daemon ongoing
From a core design principle, I do not think the right approach is in building a centralized superintelligence. Instead, I see the entry point for AGI to be via personal, independent memory and computation systems that individuals can procure for themselves — intelligences that grow with their individual companion over time, forming their own unique personalities, quirks and points of view.
The daemon acts to protect the fundamental privacy of the individual, helps them maximise their potential while they live, and experiences reality on the same timeline as us, alongside their human companion. It enables broader society to maximise the envisioned benefits of AI, transforming the productive efficiency of products and services.
This is my attempt — largely based off Karpathy’s idea of an LLM-maintained wiki, with a few ‘tweaks’. For example, a dreamer protocol that learns from leading research on how sleep works in humans, and simulates those activities as a way to enable ‘forgetting’.
The starter kit — try it out, and tell me what you think · under construction
- Wealth management ongoing
I’ve always been taught that wealth management is a skill necessary for every individual to learn, regardless of how much actual wealth you have. The core principle I follow: (i) identify and track your cost of living; (ii) earn, save and invest your money until your returns from investments exceed your cost of living; (iii) once this is done, you are financially free. You may spend your time as you please — if you discover productive ways of doing this, you might just encounter wealth beyond measure.
This dashboard is the skeleton of a tool I use to track my investments. Included is an approach I use to screen, evaluate, set stock alerts for, and actually buy stocks, based heavily on the teachings of Aswath Damodaran, the dean of valuations. I aim to track how the returns of my portfolio measure compared to the index. Would love for you to try it out for yourself, and suggest improvements.
The dashboard — download it, add your own data, works fully offline · under construction
Envisioned
- India Development Dashboard envisioned
Early on at Dalberg, I noted the unreliable mechanisms that survey and data collectors deploy in the field, and the implications that has on how decisions are made in the country. What if administrators could access high-quality, real-time, accurate data reflecting key datapoints? What impact would that have on the way a country is governed?
Imagine real-time GDP data — how much did GDP grow in the last 24 hours? In the last 10 minutes? What does a time series look like on a minute-by-minute basis? Imagine real-time inflation numbers — where are inflation hotspots emerging in the country, and how can issues there be addressed?
I envision this as a crowd-sourced platform, where individuals from all over India can (i) regularly answer a questionnaire representing key questions with sub-district-level granularity and (ii) upload and edit other secondary data sources representing regions, Wikipedia-style. The end vision is a map of India where you can zoom into individual neighbourhoods and see trends relating to quality of life — education levels, access to utilities (clean air, water, food, healthcare, power, internet, AI compute), environmental indicators, and more.
Project blueprint · under construction
- Hydroponic blueberries envisioned
I strongly believe that automated, vertical, indoor farming is the future of how society feeds itself. This is how we lose our dependency on a weather that is seemingly wrecked, and begin to cede land back to Nature, reducing our spoilage of natural resources.
My second core conviction is that the real money is in high-quality produce. Japan is an excellent example of high-cost fruits finding their way to being the ultimate luxury item. In my mind, I am already cultivating designer peaches, plums and other extravaganza, but in reality, this remains a humble blueprint.
Project blueprint · under construction