audi0slave

5 ms·

1.
Scaling AI/ML Infrastructure at Uber(uber.com)
3 points by audi0slave 2y ago|0 comments
2.
by audi0slave 4y ago
I wonder if you ever looked at sparrow paper[1] which came out Ion Stoica's lab. Its also a decentralized in nature focusing on low latency placement. [1] https://cs.stanford.edu/~matei/papers/2013/sosp_s
3.
by audi0slave 5y ago
This paper compares xgboost vs nueural nets and an ensemble [Tabular Data: Deep Learning is Not All You Need] https://arxiv.org/abs/2106.03253
4.
by audi0slave 5y ago
AFAIK, it does. Uber does too, though both of them are moving to k8s source: ex uber compute team member
5.
by audi0slave 7y ago
Reddit shows what the world is, tik tok shows the world you want to see. if you ignore this you will become the part of the problem.
6.
by audi0slave 7y ago
Sounds a lot like the Dalton highway , we had to rent a special vehicle which had a radio to do 2 way communication with the traffic and the highway wasn't paved.
7.
by audi0slave 7y ago
Pretty ironic given the same day Juul buys[1] 123 Mission for approx 400 million dollars. [1]: https://sfist.com/2019/06/18/juul-just-bought-a-28-story-som...
8.
by audi0slave 7y ago
yes, it's log scale
9.
by audi0slave 8y ago
How do you get the watch API's without etcd? or does k3s doesn't have watch API support.
10.
by audi0slave 8y ago
No tests?
11.
by audi0slave 8y ago
Have you looked at this: https://www.usenix.org/system/files/conference/atc18/atc18-o...
12.
Sentimental Versioning(sentimentalversioning.org)
1 points by audi0slave 8y ago|0 comments