About the author, brief

Scott Dobbins

MS, Biochemistry, Stanford University (high-throughput image segmentation and analysis on dividing cyanobacteria)

BA, Biochemistry, Columbia University (ion channels and such)

Data Science Bootcamp Graduate (3-month, full-time, immersive program), NYC Data Science Academy (that’s me: back-left)

Most recently: Program Director / US Office General Manager for Sunrise International Education

Lives in Brooklyn

Fluent in Mandarin

Addicted to podcasts

 

About the author, detailed

Business experience and background (LinkedIn)

Code samples (Github) (some repositories are private, and I tend to work in large chunks and commit only tested, “full thoughts” of code, which can sometimes be large)

Python or R? 

Generally, Python for image analysis, some NLP tasks, and lower-level algorithms; R otherwise. R’s data.table is better than pandas, but numpy, skimage, and sklearn are just too convenient and powerful. I also have some experience with C, Java, Perl, and MATLAB (which I used for my Master’s thesis), though it seems like everything is in Python or R these days.

Emacs or Vi?

Vim, though it doesn’t matter much.

Individual work or team work?

Team work in which each team member has authority over an aspect of the project within their area of expertise.

Coffee or tea?

Coffee on hard days, tea on easy days.

 

Ask/tell me (/correct me on) anything: scott dot n dot dobbins at gmail dot com

 

About the site’s name

A short poem theorizes on talent with these observations:

“世有伯乐，然后有千里马。千里马常有，而伯乐不常有。”

Rough translation that encapsulates the story: “Truly great horses didn’t exist until Bo Le [legendary Spring and Autumn horse tamer renowned for picking excellent horses]. Of course, great horses have always existed, it’s just that no one had yet figured out the secrets of how to identify them among the rest.”

The poem––in seven clear, concise, narrative sentences-–crystalizes important observations about the nature of talent and excellence:

• Talent is everywhere, yet difficult to find
• Some positive traits are easily mistaken for others
• Talent is easily mistaken for related negative traits
• Talent needs the right environment to thrive
• Expertise, appropriately applied, can reshape possibility space

As the world around us is reshaped under our feet by computational technologies (from basic machine learning approaches through deep neural networks running on racks’ worth of data constantly updated on-the-fly), what approaches hold promise and in what contexts? how can we identify the best approach for a given problem? when are simpler tools better than shinier ones, and when are more advanced approaches required?

These are the questions I investigate here.

Granted, this blog is largely a repository for my pet projects and professional portfolio, but this theme provides a window into what’s going through my head as I tackle the tasks you see here.