Five Lessons from a Five-city Roadshow: The SAP HANA and AWS Sagemaker ML Workshop West Coast Tour

By Tina Tang, co-founder Women in Big Data

AWS, SAP, and Womeninbigdata.org partnered for a five-city workshop tour September 4-12, through San Diego, Los Angeles, Palo Alto, Seattle, and Vancouver. The workshops played to full rooms of data scientists in every location. Signaling strong interest in learning how the technologies of both vendors can work together in the same landscape, most attendees powered through the entire five-hour workshop. Data scientists have grit!

Lesson 1: Develop a good tutorial

It may seem obvious if you’ve never done it, but the actual development and testing of a meaningful, five-hour lab can be quite time consuming to orchestrate, test, and document. The first lesson of success is to give yourself plenty of time to make a high quality tutorial. With AWS prevalence as a cloud provider and SAP’s footprint in the enterprise market, our goal with the lab was to introduce AWS builders to SAP technology on the AWS Marketplace and to inspire ideas of how to use the technologies from each vendor together.

SAP & AWS also saw an opportunity to promote the Women in Big Data community and to tap their data scientists and developers as both as instructors and participants. Data scientists and engineers from all three organizations tweaked, debugged, and iterated on the material, seeking to: “Quickly train your TensorFlow models using Amazon SageMaker, deploy them as containers in Amazon Fargate, and finally consume them from SAP HANA express edition using SAP HANA External Machine Learning Library (EML)”. You can find it here.

Lesson 2: Build a high-performance team of diverse data scientists

When you run five workshops in five different cities back to back, using sophisticated technology and serving a sophisticated audience, make sure you have a group of ten rock star data scientists from SAP, AWS, and Women in Big Data to create it, debug, refine, document, and teach it. The virtual team worked together across organizational boundaries, national boundaries, time zones, cost centers, day jobs, and reporting lines.

Lesson 3: Watch registrations carefully

If you are lucky enough to have this combination of winning partners, cutting edge tech, and eager audiences, be sure to watch your registrations carefully. AWS had to close the registration sites after 3.5 days, because we reached capacity for each venue (even taking into account an expected 50% attrition). We had an extremely high level of engagement from participants through a long day—five hours of hands-on work. AWS security had to kick us out at the end of the day!

Lesson 4: Purpose-driven goal encourages massive collaboration to happen

Each participating organization provided a massive collaborative effort from product marketing, developer relations, alliances, partner marketing, social media, diversity, business development, and data scientists from multiple divisions (including external volunteers). Every single person went above and beyond with budget contributions, working nights and weekends, and lending moral support. The enthusiastic and positive audiences in every city were a gratifying reward.

Lesson 5: Never underestimate data scientists

This is what I’ve realized about data scientists—both the ones who attended the events and the ten that helped develop, refine, correct, and teach the material:
Data scientists have superpowers of extreme persistence and resilience. They navigated and jumped over frustrating hurdles in WIFI security, setting up accounts and policies, spinning up instances, training a data set, and then taking it all down again. Not to mention the freeway traffic they faced getting to and from the events! The takeaway for anyone working in AI: you want data scientists working for you, not against you. These professionals are tough, smart, and unstoppable.

Leave a Reply

d
c