Research work, side projects, and ideas I've vibecoded with Claude.
Work-related projects.
A mock health-record web app used to assess the digital competency of frontline health workers (ASHAs, ANMs, etc.) in the Indian public health system. It mimics the look and basic functionality of a real reporting app (e.g. UWIN). During a competency survey, a health worker fills in a record for a mock patient case, and the app captures interaction metrics (time per field, edits, errors) for later analysis.
WORK IN PROGRESS. Using LLMS to assist with title abstract screening for systematic reviews. The workflow uses Covidence extract to run multi-agent title abstract screening with human in the loop.
WORK IN PROGRESS. My professor at Hopkins, Dr. Diwakar Mohan, built the sample size calculator. The distribution explorer is inspired by Matt Bognar's probability applets.
Isolation-forest-based QA pipeline for primary health-survey data.
Projects tied to academic research.
WORK IN PROGRESS. My personal quest to understand the nature of life and evolutionary biology though the lens of simulated evolution. As well as attempst to test my own hypotheses about AI and purpose.
MPH Capstone project: Framework for a gamified digital training tool to strengthen public health systems in low- and middle-income country settings.
Personal projects and experiments.
A personal news feed that summarized my day ahead from various sources. Starting with a weather report and tracking my tasks for the day, fitness routines, news topics. As well as updated to music, comics, gamesm, and more.
My nuke in the war against my ISP. An internet speed logger that runs in the background and logs internet speed to a CSV file every 5 minutes. It also generates a dfetailed report of the internet speed to show how trash my ISP's service is.
A local Python script that opens a browser tab for designing and exporting printable, label-free street maps from OpenStreetMap data — the style used in those city poster prints.
A fun little poster builder that pulls data from TMDB to generate printable posters of Movies that were adapted from books. The posters include the book cover, movie poster, and a QR code that links to the movie's TMDB page.
Personal homepage and business card. Built as a static site generator. Pulls from ORCID to populate publications. The AfterHours engine was vibe coded but I hand drew all the assets and pixel art.
WORK IN PROGRESS. A self-contained HTML dashboard, generated by an R pipeline, that visualises global development indicators on a fixed 1990–2100 time axis and runs a small integrated assessment model (IAM) coupling climate, economy, and demography. Pulls data from World Bank, OWID, UN WPP, CCKP.