yandie

9 ms·

91.
Small Changes for Big SQLite Performance Increases(whylabs.ai)
12 points by yandie 4y ago|1 comments
92.
Horizontally scale SageMaker ML Models monitoring(aws.amazon.com)
2 points by yandie 4y ago|0 comments
93.
Data Logging with Whylogs(whylabs.ai)
1 points by yandie 4y ago|0 comments
94.
Whylogs 1.0(whylogs.readthedocs.io)
6 points by yandie 4y ago|0 comments
95.
by yandie 5y ago
WhyLabs cofounder here, so my opinions are probably biased. When it comes to MLOPs, data makes it much more complex to handle. Think of it as the curse of dimensionality. Nobody wants to deal with metrics across tens, if not hundreds or tho
96.
WhyLabs, AI Observability as a Service(thesequence.substack.com)
1 points by yandie 5y ago|0 comments
97.
by yandie 5y ago
This is why we've been trying to encourage people to think about lightweight data logging as a mitigation for data quality problems. Similar to how we monitor applications with Prometheus, we should approach ML monitoring with the same
98.
Detecting Semantic Drift Within Image Data: Monitoring Context-Full Data(whylabs.ai)
2 points by yandie 5y ago|0 comments
99.
by yandie 5y ago
It's kind of a spillover effect where your customers need to meet compliance requirements from other customers such as US presence (not limited to US government btw - mostly around data handling requirements by Infosec teams). Thus as
100.
by yandie 5y ago
Compliance. If you're building a SaaS startup then it's very important to have physical presence in the US
101.
by yandie 5y ago
Blocker would be the H1B lottery. You'll have better luck using L1 and then convert to a green card than going down the H1B route.
102.
by yandie 6y ago
Firecracker is a famous one that I can think of.
103.
Data logging for your Kafka ML Pipelines(whylabs.ai)
1 points by yandie 6y ago|0 comments
104.
Data Logging: Sampling versus Profiling(whylabs.ai)
1 points by yandie 6y ago|0 comments
105.
Show HN: Streamlining Data Monitoring in MLflow(whylabs.ai)
2 points by yandie 6y ago|0 comments
106.
1.2TB of Humblebee Image Data(vimeo.com)
2 points by yandie 6y ago|0 comments
107.
by yandie 6y ago
ex-AWS here. IAM is definitely a service! It's super powerful underneath the hood. However as a caller, you need multiple levels of caching underneath the hood (all built by IAM team). That's part of the reasons why you can't
108.
Anaconda Introduced New Commercial Edition License(anaconda.com)
2 points by yandie 6y ago|0 comments
109.
by yandie 6y ago
That's what a VPN is for. Your ISP can't see the destination up addresses and only see your connection to the VPN endpoint
110.
by yandie 6y ago
Their slogan has always been "build a better internet". I find their product strategy every fascinating
111.
by yandie 6y ago
Just mean that you'll need to deal with latency, either through provisioned capacity lambda or build it into your service SLA :)
112.
by yandie 6y ago
Depends on whether you need GPU for model serving. Most companies I talk to use CPU for serving and GPU for training, so it's not a bad idea
113.
by yandie 6y ago
Lambda still needs to load your container. So there's gonna be slow start cost (but hopefully they can improve caching overtime)
114.
Amazon SageMaker Feature Store(aws.amazon.com)
5 points by yandie 6y ago|0 comments
115.
Amazon SageMaker Data Wrangler(aws.amazon.com)
2 points by yandie 6y ago|0 comments
116.
AWS Lambda – Functions with Up to 10 GB of Memory and 6 VCPUs(aws.amazon.com)
71 points by yandie 6y ago|43 comments
117.
AWS Lambda now supports container images(aws.amazon.com)
2 points by yandie 6y ago|0 comments
118.
by yandie 6y ago
CF, obviously - it's AWS! It probably will support CDK in the future (underneath is CF template), but I don't see terraform being supported unless AWS acquires the company
119.
by yandie 6y ago
Taking a look at the template repo, it seems to be just CloudFormation at the moment: https://github.com/aws-samples/aws-proton-sample-templates/t... I guess I won't be trying this until they support CDK. For
120.
by yandie 6y ago
Have you done a comparison with https://datasketches.apache.org/'s KLL algorithm?

More ›