Blogs

Virtual data science can rise to the challenge in unprecedented times

The IBM Data Science and AI Elite (DSE) offers remote engagements to help you thrive and succeed on your journey to AI

Business Director, Data Science and AI Elite, IBM

As we learn to cope at a personal level with the dynamic development of the COVID-19 outbreak, we are seeing an increasing impact to businesses in every industry. Supply chains are broken, demands are shifting and resources are shrinking. How do you prepare your business for the economic and operational impact?

The IBM Data Science and AI Elite (DSE) team was created over two years ago to work with clients in every industry to help them harness data science and AI in all aspects of the business and bring value. The current crisis creates a more urgent need to optimize resources, more accurately predict demand and assess risk. But how do you do that when your team is scattered and isolated?

Employing proven Agile AI methodology and practical experience coupled with guidance for remote engagements, the DSE team can put together key strategies you can use to achieve meaningful impact in just two to six weeks with discovery, scoping and kick-off.

Discovery: Define the problem at hand, determine the scope, identify the stakeholders

  • Design thinking. Typically, design thinking workshops are active and dynamic in-person sessions. To approximate this critical step in a virtual workplace, we created a template using Mural to fill the gap. We recommend that you find a virtual tool to guide your team through collaboration and brainstorming.
  • Whiteboarding: Whether you are drawing out the data infrastructure for your team or mapping out a business process, the digital world can offer you a whiteboard substitute. Most video conferencing tools have an annotate feature you can use, but sometimes a low-tech solution works just as well. Grab a pen and paper, draw it out hold it up to the video camera.

Scoping: Scope the sprints, get data access, project plan, document

  • Coordination: Without daily in-person interactions, being proactive can keep a project on track. Decide early on your project management tool, validate that the team is comfortable with the selection, and obtain buy-in for using it to track progress. Also, set regular check-ins at the onset of the project – and always have the next meeting scheduled before you end a call.

Kickoff: Assess team skills, set expectations, plan sprints

  • Team engagement and buy-in: It’s tempting to sit down, pull up a window on our laptop and multi-task through a meeting. The temptation to divide our attention increases once that meeting moves from a conference room to a call. The key to keeping the team engaged is to establish a “camera on” policy from day one. Set the example for your team. You don’t need to have a blank white background. You might see a child or dog running around - and you most certainly won’t be seeing any suits - but having that camera on will help hold their attention and give you a chance to regain some of the valuable non-verbal feedback from facial expressions and body language.
  • Team skills assessment and assignment: In the world of data science, skill assessment is critical for assigning the right tasks to the right team members. This is typically done when sitting side by side, collaborating on a project, and seeing where someone thrives and struggles. To adapt this experience to virtual teaming, use specific and directed questions for early assessment, then check in often to reassess. Ask team members if they have used a certain library or package before. Ask what they expected to see for results -- and how it matched against their current outputs. Communication also becomes essential for assigning out tasks. Rather than organically assigning items to a group, purposefully break down tasks for specific owners and agree on check-in dates.

Hands-on: How to conduct active coding

  • Coding and modeling: It’s tempting to hunker down and code by oneself and “knock it out.” However, our team has always believed in the power of collaborative coding. With pair-coding you are not only providing a much-needed morale boost in the time of social distancing, but also creating a venue for programmers to learn from each other, speed up on-boarding, and decrease the likelihood of bugs. Collaborative coding doesn’t have to stop just because the office building is closed. Using IBM’s Watson Studio platform, our team has conducted hundreds of successful pair programing virtually on a daily basis. Log onto a video call, share your screen and co-create the solution.
  • Daily stand-ups: Lean on agile practices such as a daily, camera-on stand-up to keep your team in sync. Ask what they accomplished yesterday, what they plan for today and what roadblocks they are hitting.

These are unprecedented times. Preparedness and planning ahead will be key to not only survive but thrive on the other side of this. The DSE team is here to partner with you so your organization can come out stronger than ever.

The DSE team is offering free weekly webinars covering topics like industry use cases, best practices and other AI-related topics. Register here.  

Also available are Industry Accelerators, including sample data, models and documentation to kick start your project so you can quickly solve your most pressing business problems.

Find out how the DSE team is taking the AI journey virtually. We will plan, co-create and prove the project with you based on our proven Agile AI methodology.

Visit: ibm.com/community/datascience/elite

Accelerate your journey to AI.