eggie5

22 ms·

241.
by eggie5 9y ago
I'm in that looksee contest the article says got works wide interest but when I look at the online dashboard it says only 556 people are registered and in the running....
242.
by eggie5 10y ago
Each company does the 5 interview format and some of them, an Apple data science group for example, ask or a presentation of your research. Random people from the team, group or larger org will interview you, seemingly to remove any bias. C
243.
by eggie5 10y ago
in my experience only 1/5 of the interviews will touch on ML
244.
by eggie5 10y ago
that's exactly true. You get interviewed by 5 different people across the org.
245.
by eggie5 10y ago
I interviewed at all the big bay companies (and the seattle one) for machine learning earlier this year and few of the questions were classical machine learning questions. Most of the questions were programming an computation.
246.
by eggie5 10y ago
I've spoken w/ them. A very interesting team inside of amazon!
247.
by eggie5 10y ago
I've spoken w/ them; it's a very interesting product!
248.
Show HN: Visualize 4096D CNN image features in 2D using t-sne(sharknado.eggie5.com)
1 points by eggie5 10y ago|0 comments
249.
by eggie5 10y ago
all of your jokes about the dashboard not turning red b/c the icon is hosted on US EAST are true: Amazon Web Services‏Verified account @awscloud 8m8 minutes ago More The dashboard not changing color is related to S3 issue. See the b
250.
by eggie5 10y ago
confirmed: dockerhub is down too
251.
by eggie5 10y ago
is this affecting dockerhub for anyone?
252.
by eggie5 10y ago
yes, confirmed.
253.
by eggie5 10y ago
This one is interesting to see the "statistical" other side of the industry vs machine-learning people. For example I don't think gradient descent is used once in that book.
254.
by eggie5 10y ago
my Pinterest spam bot revenues will take a dip :(
255.
by eggie5 10y ago
Most rec-sys these days are focused on collaborative filtering techniques based on SVD. This is the motivation behind the winning netflix challenge submission.
256.
A quick background recommender systems including the BPR algorithm(eggie5.com)
4 points by eggie5 10y ago|1 comments
257.
by eggie5 10y ago
i talked to my friend, a UPS driver, about this last year when this same story surfaced on HN. He said it's not really true and that they turn left all the time -- but he was aware of some training about it.
258.
Model Evaulation: Unbalanced Datasets(eggie5.com)
1 points by eggie5 10y ago|0 comments
259.
by eggie5 10y ago
Your describing collinearity. Yes, The PCA reconstruction components will not be correlated.
260.
by eggie5 10y ago
I know how to solve the application challenge, i've just been putting it off since December!
261.
When when building a ML model, extract the test set before you scale the data(eggie5.com)
2 points by eggie5 10y ago|0 comments
262.
by eggie5 10y ago
To counter all the negatives, I had a great experience emailing Alex directly! Responses were prompt - Great company...
263.
by eggie5 10y ago
Kubernetes is written in go
264.
A quick look at regularization in machine learning models(eggie5.com)
2 points by eggie5 10y ago|0 comments
265.
by eggie5 10y ago
eggie5.com is on k8s
266.
by eggie5 10y ago
Was working w/ Anney recently and she was helpful and very prompt!
267.
by eggie5 10y ago
I always thought some type of driver was more common...
268.
by eggie5 10y ago
Or as it's been renamed last month to: "StatefulSet" https://github.com/kubernetes/kubernetes/issues/35534 Mostly what I was commenting on was how they are building some type of declarative inf
269.
by eggie5 10y ago
A lot of what they are describing in the first half sounds like kubernetes. I wonder if they could do it again would they adopt it...?
270.
by eggie5 10y ago
Cats

More ›