Ok_Cartographer5609

joined 1 year ago
 

I have build a small project using rust to generate better help statements/instructions for CLI tools. It takes in the name of the tool and produces step by step instructions.

Feel free to have a look at the repo: https://github.com/d1pankarmedhi/ghr

I would appreciate your feedback/suggestions for improvements or feature addition :)

[–] Ok_Cartographer5609@alien.top 1 points 11 months ago

Agree. I got into my 1st ML role last year. All self-taught. I've done all sorts of work - from ETL, CV, sentiment pipeline (mostly SWE stuff) and now LLM-based Information retrieval systems. My work mostly revolves around applied ML but I do have an interest in knowing the bits and bytes of ML as well. So currently teaching myself all about transformers and lms.

But it is also true that, earlier getting started with ML was easy - no need for heavy machinery/resources. But nowadays, you will need high computing power to even get started on learning something about large language models.

 

I want to know the tools and methods you use for the observability and monitoring of your ML (LLM) performance and responses in production.

My bad. Didn't know where else to put this, so since we have such an amazing community here, I thought of posting it here.

 
  • Read papers/articles/blogs for new and updated models. (I work mostly on Language systems)
  • Build POCs using new tools/models that can solve a problem or generate a new one :)
  • Get data for this. Handle databases/data stores for storing data, on the cloud.
  • Build and deploy pipelines connecting ML services with the databases. And potentially build something which is presentable.
  • Monitor for abnormality and if there's any, republish the pipeline once resolved.