ASES was an incredible opportunity to experience the energy of Stanford campus life and SF that I've only spectated on X thus far. Some of my favourite memories are...
- Waymo to Twin Peaks at night
- Some memorable speakers were
- Prof. Russ Altman on using electronic health record embeddings to enhance phenotyping of patients.
- Michael Marks (Celesta Capital) on framing your early 20s
- Shane Gu (Google DeepMind) on the future of Reinforcement Learning
- Edward Jung (Intellectual Ventures) on inventing for a good future.
- Campfire burgers and s'mores at the firepit near Denning House
- A stroll through the sprawling Bayfront Park opposite 3rd street with Karan Singhal about his vision for OpenAI and their place in healthcare broadly.
Over the course of two months I dove into cardiology immunometabolism - the idea that the interplay between lipids and chronic inflammation that drive atherosclerosis and coronary artery disease.
The data I analysed was cytometry on time-of-flight (CyTOF) where metal-tagged antibodies that recognise specific immune cell surface markers are separated by mass cytometry.
The work led to a presentation and scholarship at the Trajectories to Heart Failure: Immunometabolic Mechanisms conference, in the beautiful mountain ranges of Keystone, Colorado!
You can see some press releases here.
& Kolling Institute
I have been blessed with two of the most incredible people Prof. Jean Yang and Prof. Gemma Figtree as mentors for my research journey. They've fostered my diverse interests and encouraged me to explore and seek out areas I'm interested in freely. I've been involved in the data analysis of many areas of study including:
- Quantifying resilience and susceptibility to coronary artery disease beyond traditional risk scoring
- Immunomodulatory mechanisms of statins with CyTOF data
- Evaluation of language models in interpreting bioinformatics analysis
- Extreme temperatures and their effects on myocardial infarction mortality. You see a press release for a travel scholarship won with this work here.
- More recently... Single cell RNA-seq
By complete accident, I stumbled into the startup scene through a conversation on Earlywork. Spent the next 7 months at Eucalyptus - a digital healthcare startup. One of the steepest learning curves for me. What I did while I was there:
- Helped build a growth engine that drove a 150% increase in organic non-paid, non-branded traffic to 20k/month
- Built a statistically robust A/B testing pipeline that considered break-even periods and business costs unlike off-the-shelf solutions such as Google Optimise
- Wrangled customer data to predict customers with high risk of churn and produce recommendations for retention optimisation
I spent the summer re-discovering tensorflow, sklearn, pytorch and fastai. Diseased blood cell counts were still manual in pathology labs. The goal was to use an object detection model to perform automatic cell counts even on images with very small cells.
- We used YOLOv5s with SAHI which allows you to slice images and sequentially inference
- The bounding boxes of WBCs were then extracted and used to crop cells out.
- These cropped cells were then passed through an image classifier.
At 19 fresh out of high school, I started MissionMed. In my journey to gaining admission to medical school there was so much knowledge that was ‘un-documented’. My philosophy is if I've thought through a decision or been through a process, it should be documented somewhere for myself and others to access.
All the existing incumbents were unscalable service-based consultant business models. I knew I could productise the experience, standardise the quality AND with better margins.
Documenting my processes made it easy to hire contractors from Upwork to gradually offload my work. It still goes strong today.
- Wrote my first book in Overleaf LaTeX
- Convinced Dr. Jason Yu to jump on board
- Directed and produced a whole video curriculum (I learned Premiere Pro and EpicLightMedia's youtube channel in the process)
- Built the entire LMS in Wordpress & Learndash and custom HTML, JQuery, PHP, CSS
- Grassroots growth-hacking, articles targeting long-tail SEO keywords and webinars to parents to generate leads
- Wrote Python scripts to automate administrative tasks.
I began teaching Physics and acquired some of the most useful interpersonal skills here - still do. Three main learnings:
- First principles framework - made me generative rather than replicative
- Precision of word choice - increasing information density of communication
- Explanation toolkit - analogies, recognising reasoning fallacies
I built a site and designed an entire 4-term, 40 week Physics curriculum which you can find on this website.
Tragically, it's valiant run was cut short by recent Google Play Developer Program policy changes. Funnily, the
Now of course, PayPal is not for under 18s, so I convinced someone on a forum call thebot.net to pay me in Amazon gift cards and bitcoin after I transferred my Fiverr balance to their PayPal account. This was my first foray into cryptocurrency and I had something close to 2 bitcoins which was about $700-$800 at the time.