paint-brush
⛓ Check the first ML Value Chain Landscape shaped by ML practitioners!by@kseniase
606 reads
606 reads

⛓ Check the first ML Value Chain Landscape shaped by ML practitioners!

by Ksenia SeOctober 13th, 2022
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

TheSequence reached out to all our readers (over 144,000!) to help us create a qualitatively new ML Value Chain Landscape. Half of those involved in ML are struggling with data processing and model monitoring. Currently, no player on the market covers the entire ML value chain – we’re yet to come up with an all-encompassing ecosystem with well-defined but interconnected parts, offering a flexible work environment that could effectively meet all ML needs. We hope to gain more valuable insights for future iterations of the project.

Companies Mentioned

Mention Thumbnail
Mention Thumbnail

Coin Mentioned

Mention Thumbnail
featured image - ⛓ Check the first ML Value Chain Landscape shaped by ML practitioners!
Ksenia Se HackerNoon profile picture

Recently at TheSequence, we reached out to all our readers (over 144,000!) to help us create a qualitatively new ML Value Chain Landscape.


This month-long process involved getting as much existing research together as we could get our hands on (from CB Insights, MAD, Gartner, and top media outlets among other sources), organizing it in a way that could be understood by everyone, and then reshaping it together in the best possible way based on each member’s personal experience with different ML stages.


Thank you for participating! Now you can see the result of that huge undertaking – the final version of our ML Value Chain Landscape:


We chose color codes to indicate which vendors cover 2 and more stages of the ML Value Chain.


Among some of the most revealing insights we’ve been able to gather in the process are the following:

  1. Half of those involved in ML are struggling with data processing and model monitoring.
  2. Most solutions aren’t optimized for later stages during development. Model monitoring in particular involves laborious work (problems pile up, often with no end in sight).
  3. Data processing is largely a disjointed enterprise – there’s no platform with insufficient user-friendliness and interoperability adding to the frustration. A solution that’s “easy to use, easy to configure, and easy to scale” is hard to come by.
  4. Little interaction between different stages is a real obstacle to project completion. Many techniques and software tools don’t agree with each other; at the same time, there’s little room for collaboration between professionals with different backgrounds and varying levels of expertise.
  5. Currently, no player on the market covers the entire ML value chain – we’re yet to come up with an all-encompassing ecosystem with well-defined but interconnected parts, offering a flexible work environment that could effectively meet all ML needs. So far only Scale AI, Toloka AI, Appen, Abacus AI, Vertex AI have been able to cover 5 out of 6 stages. Is it possible or does it even make sense to cover all six stages? We are sure we will find out soon, as the ML industry is developing extremely fast with so many amazing startups working on it.


Although this is now a finished project as far as our community surveys go, we need to remember that landscapes evolve continuously alongside emerging technology and its applications. With that in mind, please share our ML Value Chain Landscape to get as much feedback as possible from more data scientists and ML practitioners around the globe. We hope to gain more valuable insights for future iterations of the project!


Also published here.