Hey Byron, thanks for the post and the fantastic summary on MLOps. Technology choices are often driven by people & processes, and that's what makes it so difficult to make tech stack decisions.
You mentioned that Feature Store + Kubeflow = 90% ideal MLOps Platform. What would the remaining 10% be? Fairing? Responsible AI including ethics and bias checks? Experiment design?
Two years ago, I earned my badge of dishonour by leading an inefficient operation. I come from a data science & machine learning background, so of course, we’ve learnt about DevOps from our engineering colleagues.
Or so we thought.
It was inexplicable to us at the time, as our data scientists were sitting right next to the data engineers. We followed all the good agile practice of daily stand-ups, talk about our blockers, and didn’t have the “throw-it-over-the-wall” attitude. We collaborated closely, and our scientists and engineers love each other. …
There have been many articles focusing on how machine learning can and is helping during the pandemic. South Korean and Taiwanese governments successfully demonstrated how they used AI to slow the spread of COVID-19. French tech company identifying hot spots where masks are not being worn, advising where governments can focus on education. Data and medical professionals collaborating to index medical journal papers on COVID-19.
Participation from the data science community is inspiring and the results are outstanding.
But this post is about the opposite. I am not here to tout the successes of machine learning. I am here to…
As a data scientist, you expect to get a job that lets you do cool stuff — Big data, Big machine (or cloud, like the grown-ups), and Deep neural networks. Reality quickly creeps in as you realize the mismatch between your model, your project manager’s timeline, and your stakeholder’s expectation.
Your project is creeping further and further from the original scope, and you are miles away from completion. The demo is Tuesday, your eyes are watering and palms sweating.
“That’s not in the requirement!” — almost every developer
Previously, I wrote about the importance of defining data science scope with…
As one of the fastest-growing industries, being a data scientist without a research group can be incapacitating. At the end of last year, I realized I am often missing out on new development in the industry, unknowingly reinventing the wheel, and falling short in conversations with experts. Being a data science consultant, my stress-level exploded as I constantly felt under-prepared going into client meetings.
So as my goal for 2020, I decided to set myself free of this stress.
“To know that you do not know is best. To think you know when you do not is a disease. Recognizing…
As a data scientist, you expect to get a job that lets you do cool stuff — Big data, Big machine (or cloud, like the grown-ups) and Deep neural networks. Reality quickly creeps in as you realise the mismatch between your model, your project manager’s timeline and your stakeholder’s expectation. What they needed (often) is not a 128-layer ResNet, but a simple select & group by query that delivers actionable insights.
There you are two months into your new job with your shiny model that just got shelved, grunting: “What is actionable insights anyway. Insights for who? Actioned with what?”
In April this year, instead of organising my 20-something birthday party, I spent hours writing motivating emails to my company to convince them to send me to Deep Learning Indaba*.
My very generous employer instantly grant my wishes. But once my wish was granted I took an entire weekend to draft the answers for the conference application on Google Form. The application form was short — but after being rejected last year I really had to beef up my game.
The competition is high. You will have to demonstrate real passion for Machine Learning to be selected.
I would do…