#TurnTheLens #MattRocklin #Coiled #JeffFrick
Matt Rocklin has been in the center of the Python renaissance in enterprise data science, at Anaconda, NVIDIA, and now Coiled, Matt's start up focused on helping enterprises more easily deploy parallelized python in production at scale. Artificial Intelligence (AI) & Machine Learning (ML) are two of the most important technologies of our time, data scientists write the algorithms that make it possible. Coiled is enabling more data scientists to do their job better by removing friction, both for scaling (to cloud) and working within the confines of corporate governance, compliance, and economic requirements. In this far ranging interview, Matt discussed the continuing advances in data science, open source in the enterprise, the python resurgence, and the challenges of building a company during a global pandemic.
Thanks Matt.
--------------------------------------------------------------------------------------------------------------------
00:00 Intro
00:29 Introducing Matt Rocklin, Coiled
01:16 Organizations are sitting on treasure troves of data, but gaining insight and value from it is hard
01:26 Data scientists pour through the data to construct new policies, operational systems, & algorithms
01:49 Python is the standard data science, data engineering, machine learning language today
02:02 Matt, open source maintainer for NumPy, Pandas, Scikit-learn, Jupyter, etc.
02:09 Focus on the problem of scaling up the existing community to work on larger data sets
03:11 Open source as a force multiplier, a way to help others be more productive
03:27 Dask origin story
03:38 People love Python, but when they go from a gigabyte to large data sets, everything breaks
03:53 Dask was designed to fix the disconnect, to scale the existing ecosystem of tools
04:03 Dask has breathed new life into the Python space, opening up completely different problem sets
04:51 Matt compares CPUs to GPUs in machine learning
05:28 NVIDIA RAPIDS, Dask and Changing the economics of AI to open up the problem set
05:56 NVIDIA RAPIDS, scale, not one GPU but hundreds
06:50 Parallel also means less power, lower costs, opens up a different set of problems
06:55 1,000x speed up for only a 10x bump in price
07:00 Better economics opens up a broader set of problems
09:38 Coiled origin story
09:38 Coiled was designed to make scaling Python easy in the cloud, most commonly for data science and machine learning
09:47 Coiled is based around the open source technology Dask, which is the pre-eminent Python native solution for distributed computing
10:23 People love Dask, but elevating it within an organizational is hard, often for DevOps of organizational challenges
11:12 Coiled offers a hosted and managed Dask offering in the cloud
11:19 A consistent challenge for data science teams, easily moving from desktop to cloud scale
11:55 Coiled solves the problems IT cares about, security, user management, cost management, et.
12:46 IT has a mandate to support scalable machine learning platforms, and Coiled fits their needs
12:54 IT likes Coiled because we solve one problem really well, and get out of the way for everything else
13:10 Coiled is also designed for the individual user, designed to make it easy for anyone to scale out Python anywhere
14:09 Adding new data sets to algorithms
14:51 By allowing yourself to hit the full date fire hose, you can make much more accurate models
16:51 By building applications for the largest organizations in the world, we're also able to serve broader public
16:57 Our Mission, providing computation and accessibility for everybody
17:36 Matt on modern data tooling changing the way we thing about analytics
17:53 Everything is broken apart so you can look at different kinds of applications and technologies
18:32 Open data
19:24 Impact of cloud infrastructure
21:21 Matt on the challenges in moving from developer to CEO
21:27 Chess and the CEO
26:24 Matt on open source
------------------------------------------
Disclaimer and Disclosure*
DISCLOSURE*: This interview was sponsored by Coiled.
Neither Coiled nor other sponsors have editorial control over the content.

Quotations are attributed to the original authors and sources.

All products, product names, companies, logos, names, brands, service names, trademarks, and registered trademarks (collectively, *identifiers) are the property of their respective owners. All *identifiers used are for identification purposes only. Use of these *identifiers does not imply endorsement. Other trademarks and trade names may be used in this document to refer to either the entities claiming the marks and/or names of their products and are the property of their respective owners.

We disclaim proprietary interest in the marks and names of others. No representation is made or warranty given as to their content. User assumes all risks of use.