This is great. I appreciate the specific examples of where DE can leverage GenAI. I subscribed based on this and hope you will go deeper on this GenI + DE topic.
Great article, thanks for writing and sharing! Member of your v4 DE bootcamp here and interested in your AE bootcamp.
Just wondering, are you personally more interested in business/analyst archetype of DE over the software/infra one? I imagine the latter would be a backend/devops/DE hybrid. My understanding was that at Netflix (where you did your most technical work), you were the software/infra flavor or DE.
Are you just more interested in the analyst archetype now or do you believe there will be more demand for it overall?
I personally like the engineering archetype more than the business side.
Theres more money in teaching analytics people engineering than there is in teaching engineering people analytics. Software engineers aren’t nearly as motivated to learn from me as analytics people are. I can tell my the short sales window I did for this version of the boot camp that I was right about this too.
And the analytics boot camp is less error prone. Doing live coding demos in front of 250 people is stressful. When it’s more just SQL and Python, I can deliver on value there without mistakes. I have to rehearse the engineering content more which makes it much lower ROI on my time
Makes sense! I image that's because there are more people in analytics overall and that it's easier to go from technical -> business than the other way.
I'd love to know what you would put in a curriculum for the engineering archetype. Maybe you can put that in one of your next blog posts!
I am not an expert, but what my initial thought is, we will still be requiring developers to write original codes to solve unique problems. Because any generative models just tries to mimic a distribution function of its train set(as far as I have studied till now).
So any code generate by gpt or any generative model is just combining multiple suitable code snippets(modifying them a bit based in the input context given by user) written by some developer/(s).
So at the end of the day, gpt like models can help to maintain the pipelines and write the code for some generic problem. But for more specific unique type of problem will be requiring developers intervention. It will just reduce a lions share of the task though.
This is great. I appreciate the specific examples of where DE can leverage GenAI. I subscribed based on this and hope you will go deeper on this GenI + DE topic.
Great article, thanks for writing and sharing! Member of your v4 DE bootcamp here and interested in your AE bootcamp.
Just wondering, are you personally more interested in business/analyst archetype of DE over the software/infra one? I imagine the latter would be a backend/devops/DE hybrid. My understanding was that at Netflix (where you did your most technical work), you were the software/infra flavor or DE.
Are you just more interested in the analyst archetype now or do you believe there will be more demand for it overall?
Thanks!
I personally like the engineering archetype more than the business side.
Theres more money in teaching analytics people engineering than there is in teaching engineering people analytics. Software engineers aren’t nearly as motivated to learn from me as analytics people are. I can tell my the short sales window I did for this version of the boot camp that I was right about this too.
And the analytics boot camp is less error prone. Doing live coding demos in front of 250 people is stressful. When it’s more just SQL and Python, I can deliver on value there without mistakes. I have to rehearse the engineering content more which makes it much lower ROI on my time
Makes sense! I image that's because there are more people in analytics overall and that it's easier to go from technical -> business than the other way.
I'd love to know what you would put in a curriculum for the engineering archetype. Maybe you can put that in one of your next blog posts!
I am not an expert, but what my initial thought is, we will still be requiring developers to write original codes to solve unique problems. Because any generative models just tries to mimic a distribution function of its train set(as far as I have studied till now).
So any code generate by gpt or any generative model is just combining multiple suitable code snippets(modifying them a bit based in the input context given by user) written by some developer/(s).
So at the end of the day, gpt like models can help to maintain the pipelines and write the code for some generic problem. But for more specific unique type of problem will be requiring developers intervention. It will just reduce a lions share of the task though.