15 Tips to Scale a Large AI/ML Workshop - Both Online and In-Person
10 likes2,618 views
The document provides 15 practical tips for scaling workshops effectively, emphasizing the importance of collaboration, content reuse, and structured Q&A channels. It advises on cost management, testing across regions, and the implementation of a backup plan to ensure smooth delivery. The tips also encourage leveraging feedback for improvement and maximizing the impact of workshops through strategic content sharing.
15 Tips to Scale a Large AI/ML Workshop - Both Online and In-Person
- 1. 15 Tips to Scale a Large Workshop Chris Fregly (US) @cfregly Antje Barth (DE) @anbarth AI and Machine Learning Specialists https://datascienceonaws.com
- 2. No. 1 Find a trust worthy buddy. Alternate between speaking & live question-answering.
- 3. No. 2 Create your own fly wheel. Books + Online Trainings 1-hr Webinar Meetups + Conferences 8-hr Workshop Don’t rely solely on marketing team. Find your own channels. Plus 50,000+ Members Across Many Meetups Worldwide Millions of Members Across the Globe Reuse Content!
- 4. No. 3 Minimize the Q&A Channels. Stick with a single channel for Q&A during an event.
- 5. No. 4 Keep everything “on rails.” Minimize degrees of freedom and maintain sanity.
- 6. No. 5 “Always Be Testing” “Hook people on random rewards.” --Nir Eyal “We never play the same show twice” --Grateful Dead
- 7. No. 6 “If you can’t measure, you can’t improve.” --Peter Drucker “See where people walk, then build the sidewalk.” --Chris’ Dad
- 8. No. 7 Don’t let negative feedback discourage you. At scale, you will always get negative feedback. Aggregate, evaluate, and improve if consistent.
- 9. No. 8 Take control of the sandbox. Use clean-room environments, always provision extras. Do not run on attendee’s laptop or cloud account.
- 10. No. 9 Know the cost of the sandbox. Everyone will ask. Stay under $25 for full-day events.
- 11. No. 10 Test in all regions. Total capacity, different instance types/limits, API bugs. Chris’ actual shirt from reInvent 2019
- 12. No. 11 Have a backup plan. git pull Make it easy to fix bugs on the fly, if needed.
- 13. No. 12 Attendees miss instructions. Use programatic safeguards vs. text instructions.
- 14. No. 13 Update the workshop immediately after delivery. git push Fix whatever was non-optimal ... before you forget.
- 15. No. 14 Share and maximize ripple effect. Our content has enabled 5 new community workshops. We created over 30 Github issues to improve SageMaker. Chris’ actual ukulele and Grateful Dead song book.
- 16. No. 15 Use FAQ chatbot to handle commonly-repeated questions. Focus on high-value questions, automate the rest. Will this session be recorded? Recordings are posted to our YouTube channel. Where? Yes, this is recorded. Our actual FAQ chatbot with Amazon Lex + Kendra
- 17. Summary Maintain Control and Minimize Degrees of Freedom Maximize Impact and Increase Ripple Effect Hook People with New Content Create Your Own Flywheel Chris Fregly Developer Advocate AI and Machine Learning Specialist San Francisco, USA @cfregly https://linkedin.com/in/cfregly Antje Barth Developer Advocate AI and Machine Learning Specialist Dusseldorf, Germany @anbarth https://linkedin.com/in/antje-barthhttps://datascienceonaws.com