Meet Dave Crumbacher 4

David Crumbacher is the Tech Lead for the Lilly Clinical Open Innovation Team, giving direction and guidance to a group of developers that, in his words, are “gifted, unique, and highly valued.”  Dave’s history at Eli Lilly and Company on the forefront of transformational technology is deep.  Dave joined Lilly after graduating from Southern Illinois University with a degree in computer science.  His contributions at Lilly have helped drug developers leverage leading technologies to innovate in the fight against disease.

Dave brings more than 20 years of experience in high performance computing and networking to Lilly Clinical Open Innovation.  From clustered VAX environments to today’s Internet-enabled cloud computing, Dave’s career has focused on helping scientists take advantage of computational and networking power.  He has a unique capacity for translation that few people have; he has a gift to share the  most complex technology with those who need to apply it.  It is a gift he attributes to patient coaching from his earliest days at Lilly.  But it may go further back than that.

You have a unique ability to describe the most technical aspects of your work for people.  Where did you get that?

I grew up in a small mid-western town where my dad ran a TV repair shop.  In those days, TV technology was a mystery to a lot of people, but my father was the kind of man who took the time to explain to his customers what was going on with their set, and what he could do to fix it for them.  I think that his ability to do that must have rubbed off on me at an early age.

It was reinforced during some of the first discovery groups I was in during my Lilly internship during college.  I was exposed to some excellent coaching, and I really did learn early on that if your career path is in technology but you’re working for  company whose focus is not to deliver technology, it’s very important to be able to translate the value for people.

How did you begin applying your computer science degree at Lilly?

In the 80’s and 90’s, major pharmaceutical companies ran simulation programs that assisted scientists in the drug discovery process.  The VAX clusters were among the most powerful for the day, but the simulations could still take weeks to produce results.  The capabilities that I brought with me into that role improved life for the scientists by radically reducing the simulation times.  We started using a Cray supercomputer, and began to  understand what truly high-performance computing could enable.  It was intimidating for me, though, because many of the scientists were experts both in their scientific field (chemistry, biology, etc) and computational science.  I remember one of the senior computational chemists was quite happy with the improvements I made to their processes.  Soon, chemists were able to create visual molecular models and begin to design drugs from the computer.

So this was ground-breaking technology that led to some big changes for Lilly?

Making such big improvements to these molecular modeling processes showed the organization that computing power can impact drug development. And it led to Lilly making a significant investment and becoming the first pharma company in history to own and operate a Cray supercomputer.  Prior to that, scientists booked time at the Supercomputer Centers funded by the National Science Foundation like the ones at Carnegie-Mellon and the University of Minnesota.  My team made use of supercomputers at the National Center for Supercomputing Applications at the University of Illinois, Champaign-Urbana to run our simulations, because it was closest to Indianapolis.  It was exciting to be at the genesis of that.  We set up dedicated networks, trained the Lilly scientists on how to use these machines, and explored ways that Lilly could deliver scientific value to molecular modeling. 

Where did the momentum of these changes lead you?

The computing world was evolving exponentially, and it seemed I was in the middle of it at Lilly.  Our main-frame-centric organization acquired a smattering of UNIX-based machines because there just wasn’t enough power at the desktop.  Our group that led the supercomputing initiative spawned Lilly’s commitment to Unix and the inception of high-performance network computing.  Formalizing a UNIX initiative was a good example of our commitment to make this incredible computing power more broadly available and usable.  We began to discover the advantages of networked, distributed computing power.

Around that time, I had the opportunity to work with another group focused on medical informatics.  Even though it was a relatively short-lived program, we took informatics principles and applied them to the healthcare space, which was a new idea.  We were discovering the benefits of modeling techniques and simulations in the delivery of healthcare, in medical decision-making, and in the cost/benefit analysis of treatments.  We thought of using proprietary networks to connect physicians, hospitals, labs, and other facilities within the healthcare community.  But without standards, scaling and sustainability made the idea impractical for the time.

How did you overcome that?

The key to solving those connectivity issues was taking shape in the form of something called the Internet.  The Internet, as our team considered it, would be an ideal platform for fueling medical informatics, a great platform for new ways of communicating and sharing data.  The idea of an open, standards-based network leveraged for medical informatics was exciting.

We did some experimentation, stood up a Sun box, convinced the telecom guys to open the taps and on November 4, 1994, “” went live.  It was a cool moment.  It really was less about medical informatics than it was about Lilly having an online presence.  From there, we developed the ‘Internet Services Group’ where I helped to interpret the basics of using the public Internet to our business integration, marketing, legal, and corporate communications personnel.

There was a lot to be done. In addition to technical leadership, I was on point to help address legal and marketing considerations as well.  I spent a lot of time with senior executives and lawyers to shape Lilly’s Internet presence, to both establish and protect the Lilly brand.  At a later time I helped to acuire a number of Lilly international and brand domain names.

The Internet had incredible potential, but at the time, people didn’t know quite what to make of it.

How did you become involved with the Lilly COI Team?

I began working on an effort to improve the collection and handling of clinical development data.  I designed and built a proof-of-concept lab for experimenting with new technologies and applying the learning to improve the success of projects.  With its new facility, development team, and networking capabilities, we focused on bringing all of the technology together to make the clinical development process work. In a way, for me, it was a lot like those early days, bringing new technological tools to bear on seemingly unsolvable business problems.

This led to Lilly Clinical Open Innovation, where instead of Lilly focusing on improving our internal clinical development processes, we’re focused on involving the crowd in improving clinical development for everyone – a rising tide raises all boats approach.  We believe there is a perfect storm where there is a great need for transforming drug development, combined with availability of unprecedented computational power and open source software in a networked world.

You have strong principles that you live by.  Which of those do you think impact your team’s success the most?

I have discovered that, whether at home with my family, or on a team like this one, I am driven by a desire for harmony.  Harmony, for me, includes things like respecting each individual, honoring the expertise of my team members, and valuing their gifts.  These guys are unique and have complementary skills.  They all need to know that they are needed on this team.

Is there something in the Lilly COI development efforts thus far that you’re particularly proud of?

I’m proud of how we’re embracing agile development and paired programming.  It took a while to get a rhythm, but with its built-in reviews and innate collaboration, we’re seeing high quality and rapid results.   We didn’t realize how successful that would be when we first started implementing it.  And swarming, too.  The success we have when 5 or 6 developers come together for rapid progress is motivating.  Harmony means that everyone in a collaborative effort feels a real sense of value, and that relates directly to our team’s commitment to an open source environment, which is at the heart of the Open Clinical Intelligence Network model.

Moving toward a true Open Clinical Intelligence Network, what tools are you using and why?

Designing and implementing solutions for this model means we must have the same goal technically.  We use tools like Python, Django, JavaScript, JSON and the Amazon cloud.  We avoid proprietary solutions.  Instead, we leverage the open source community and these open standards, exposing the excellent work that has already been accomplished by other people, and allowing others to more easily contribute to our efforts.  Our choice to use Amazon Web Services for hosting was driven by the fact that our team is small and needs to be agile when deploying solutions.  With it, we can quickly spin up, tear down, and reconfigure the infrastructure we need without worrying about the details of managing a data center.

Our aim is to not just be consumers of open technology, but also contributors to it.  This ecosystem of “give and get” fuels innovation in the open source community.  We look forward to doing this in the open, in an unprecedented way to drive forward drug development.  We look forward to engaging with a community of developers to share and learn best practices, ideas and technology – all to make a difference in the fight against disease.


  1. Pingback: Reflections on Datapalooza « lillycoi

  2. Pingback: Pharma and Health Datapalooza IV « lillycoi

  3. Pingback: 2013 Open Source Convention « lillycoi

  4. Pingback: New R&D Model, Open Source? | bit.Med

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s