Managing Data Science Projects

34 views
Skip to first unread message

Chang Lee

unread,
May 12, 2020, 10:01:09 PM5/12/20
to Penny University

You can see the recording here: https://youtu.be/O3xGnVGUdTo


Participants


  • Mary van Valkenburg

  • Scott Burns

  • Alex Antoninson

  • Kim Alexander

  • Jeremy Jordan

  • John Berryman

  • Matthew Cronin


The Problem I'm facing


I have been leading a computer vision project of 3 people that goes head-to-head against vendor offerings. One difficulty I have as a project lead was that there are situations that I have no idea how to handle. For example, when I expect a teammate to finish the work in 4 days, but at the end of the period, the teammate would say “I’m almost done.” As far as I can tell, I have the choice of


  • Cut the work and move to another high priority task

  • Let the person work and finish the work. But this creates blockers for the team.


How do I deal with situations like this, where I can see problems are likely communication but I don’t have an answer for?


Give people ownership


Let people own their work and provide transparency into the meaning of the work, instead of assigning people tasks.


Chang: Honestly I’m still having trouble understanding what does that have to do with making a product ship on time. I can see that being a part of building a good team though.


Add regular checks to progress


What I learned was that there is a need for regular checks -- the smaller the work chunks are and the more frequent are the checks, the less likely the output will deviate from the desired outcome. 


I also have trouble breaking work down to meaningful chunks. What do I do?


Research tickets are difficult but there are ways that may help manage around it:


  • “Don’t spend more than 4 hours on this” - timeboxing when setting expectations

  • Regular presentation after research to avoid doing things that don’t matter to the team


How do I encourage people to ask more questions when I know that will save them a lot of time?


Thoughts


Oversharing own mistakes creates a culture of trust and learning.


A difficulty in management is to allow space for individual contributors but keep track of progress at the same time.


Reply all
Reply to author
Forward
0 new messages