Samuel Taylor retweeted
Who decided we should set resolutions in the dead of winter rather than the start of spring?
Samuel Taylor retweeted
Replying to @mayankmujumdar
Go fast
Be right a lot
Show your work
Trust your stakeholders
Dig deeper than asked
Bias towards action
Memorize as many metrics as possible
Communicate early and often
Find issues others cant
Samuel Taylor retweeted
That feeling when you finally get that piece of code to run that you have spent hours or even days working on.
Samuel Taylor retweeted
When is it time to deploy an ML model? Thresholding error on held-out data is a start, but not enough. The next step is writing "ML unit tests" to check performance on sub-slices.
We're entering the era of test-driven ML development:
mlinproduction.com/testing-m…
Samuel Taylor retweeted
Replying to @koehrsen_will
At my first software engineering internship, my mentor gave a stand up update that I still think about: (paraphrase)
"I'm not sure what to name this new service, which makes me think I don't understand it well enough".
Samuel Taylor retweeted
Replying to @peteskomoroch @fiddlerlabs
My reaction to 0.94 AUC 3 years ago: "Awesome!"
My reaction now: "now hold up a sec"
Samuel Taylor retweeted
We're excited to have Samuel Taylor (@SamuelDataT), Data Scientist for @indeed, presenting at #AllThingsOpen! allthingsopen.org/speakers/s…
Samuel Taylor retweeted
Sure, your #machinelearning produces accurate results. But is it useful? @indeed data scientist @SamuelDataT warns how #classimbalance can take its toll. Find out how you can detect imbalance and how much is too costly.
@ODSC #ODSCWest
go.indeed.com/imbalance
Samuel Taylor retweeted
.@Indeed Data Scientists @l3link and @atslashhashtag were at #ICLR2019 earlier in May where Ben spoke about improving job seeker-employer match models. Watch:
slideslive.com/38915705/impr…
Samuel Taylor retweeted
Treat your tech stack, skills, and career as an investment portfolio
There are options that are safe and provide a foundation, things that look good but aren't, and things that are high risk, wild ass guesses. You need to distribute time/risk accordingly
Replying to @taylorotwell
We need the zealots, man. Even when they're fully immersed in whatever idea they are pushing, they are creating excitement and buzz that gets others experimenting.
What we need more of is people preaching how to maintain a balance between experimental tech and business value.
Samuel Taylor retweeted
#Machinelearning is surrounded by so much hype it can seem like magic. Learn the math behind the magic in this machine learning overview, presented by @SamuelDataT of @Indeed at #AnacondaCON 2018. bit.ly/2HEkSql
Samuel Taylor retweeted
Excited to announce our next round of added speakers for #AnacondaCON! Take a look at all our speakers and be sure to register & take advantage of our Early Bird 2 for 1 special: anacondacon.io/speakers/ @SamuelDataT @therriaultphd
Samuel Taylor retweeted
New video from @DataSciCon Atlanta 2017: "Use #MachineLearning to Find Your Next Job" by @SamuelDataT
recallact.com/ZZD
Samuel Taylor retweeted
How I Hacked My University's Registration System with #Python and @Twilio. Nice post @SamuelDataT twilio.com/blog/2017/06/hack…














