carlosf

13 ms·

91.
by carlosf 6y ago
MLOps is just a buzzword, but maintaining ML projects in large organizations has several major pains, e.g.: http://papers.nips.cc/paper/5656-hidden-technical-debt-in-ma... I use MLFlow at my company for solving two spe
92.
by carlosf 6y ago
I wouldn't call Roomba old fashioned. It's based on emergent behavior instead of forecasting and complex rules. I haven't touched this area since 2016, but I remember coding swarms of of robots that could manage to do very co
93.
by carlosf 6y ago
Would you mind sharing the name of your product?
94.
by carlosf 6y ago
You can archive (2) with VS Code by having a development container defined in .devcontainer/devcontainer.json + .devcontainer/Dockerfile + Remote-Containers plugin. I've been doing that in my company and it's awesome to
95.
by carlosf 6y ago
You might even call it a Data Lake and get away with it.
96.
by carlosf 6y ago
- be efficient at doing stuff - automate stuff - avoid time sinks - be pragmatic about technology choices This isn't article isn't about currency, btw.
97.
by carlosf 7y ago
> How do you check your code works? When I think about my projects "working", I always try to answer the following questions: 1) Is my code doing what I believe it should be doing? That question always have objective answers an
98.
by carlosf 7y ago
Yeah, forgot about that. I also use Terragrunt and also Terratest...
99.
by carlosf 7y ago
Databricks is easily the best spent money in my stack. I also love Hashicorp's products, but unfortunately I'm not paying them anything. Before Terraform Cloud their enterprise offerings were quite bad IMO (no SaaS, had to talk wi
100.
by carlosf 7y ago
Went from Flask + FlaskRESTPlus to FastAPI. Can't recommend FastAPI enough. https://github.com/tiangolo/fastapi
101.
Ask HN: Third-Party Security Audits
1 points by carlosf 7y ago|1 comments
102.
by carlosf 7y ago
Good article! I actually like systemd, but nowadays I generally run stuff as containers, so there is less and less opportunity to use it. Here is my current setup for Airflow: - 1 container for the webserver - 1 container for the scheduler