Building a Data Science Startup

Summary of video at YouTube channel: ‘Talk Python’

Photo by Marvin Meyer on Unsplash

Original Source

Talk Python: Building a Data Science Startup (YouTube, 1 hour 12 minutes )

Jottings

• Theme: Building a Data Science Startup
• Why startup? Sick of academics (writing proposals!!)
• Decide Plan B as its Scary/Risky to start company and failure hurts
• Need to be right most times. Fallout of wrong decision should be bearable.
• Do only one thing than do consultancy to get cash and build products in parallel
• Cannot go without making money for long. Profit is best KPI.
• Side business not making money is Hobby.
• Open source == Business?:
  - 50%–50% (some open, some commercial)
  - 100% open source makes investor, employees nervous.
  - Core open source, allied services or support paid.
• New areas/ideas:
  - ML Ops (PoC->Product)
  - Toy pipelines (Quick full demo)
  - Service integration (eg. spaCy + Coiled)

The above summary was also posted in the Comments section of the original YouTube video.

More about the author of the summary, by clicking image below: