Recent Posts

More Posts

A large problem with artificial intelligence today lies in the lack of reliable data sources. To be successful, many companies have had to rely on having exclusive access to data. This works fine for large companies with clout like Google and Facebook, but puts small and medium size businesses at a huge disadvantage. Bottos is solving this problem with a blockchain solution for data sharing as well as AI model sharing.


Human behavior can be: 1. **measured** using sensors 2. **quantified** using states 3. **analyzed** using machine learning In the past, behavior has been quantified in primitive ways. Despite lack of quantification, many diseases progress due to behavioral factors. Cancer costs $$ for the health care system, but many health outcomes are driven by behavior. * Stress costs $190B/yr * Smoking cost $326B/yr [leading cause of premature death] * Substance abuse costs $100B+/yr These things aren’t just expensive, but there is generally a HUGE desire in people that smoke, abuse drugs, etc to change their behavior.


In the first year of medical school (2012), a professor told us about fecal transplants. The entire class laughed, simultaneously dismissing the idea and making inappropriate jokes. Fast forward a few years, fecal transplants are being performed more, and patients’ microbiomes are being sequenced. Research has linked the human gut microbiome to diseases such as systemic lupus erythematosis (clinical trial). The interest in the microbiome is exploding. Meanwhile, genomics interest has decreased rapidly since the GWAS bubble, and stayed relatively constant.


Price target: $120 Time frame: 1-2 years A year after my original call (http://, I am still long GILD. The positive drivers clearly outweigh the overblown fears. Positive Drivers: the recent Kite Pharma acquisition Gilead will continue to print money with non-alcoholic steatohepatitis (NASH) drugs Overblown Fears: “the hepatitis C (HCV) revenues are decreasing!” “the HCV patents are expiring!” The overblown fears should be non-issues.


Overall investment thesis to bet on: research (R&D) - heavy businesses teams with strong domain expertise going after unique use cases where AI plays a critical role in solving them But really, theres potential for all flavors of startups: Algorithm developers: new algorithms are typically very short-lived. If you serach ArXiv, hundreds of papers describing new techniques are published daily. That being said, there are various examples of successess, including DeepMind and Whetlab.



Open Science Organization

Open Science Organization - the blockchain solution for fair, transparent, quality science

Biohacker Guide

I was an editor and writer for the Biohacker Guide, hosted on Nootrobox’s website.

Deep Learning for Medicine

Deep learning for medicine: readmission prediction, early disease detection

Predictive Modeling System

Predictive modeling pipeline for healthcare. Applications for readmission prediction.


Wearable head-up display for image-guided surgery.

Selected Publications

Recent Publications

Recent & Upcoming Talks


I am a teaching instructor for the following courses at GT:

  • CSE 6250: Healthcare Analytics (2x)
  • VIP team: Predictive Health