Just gave the amazon OA for SDE 1
The first DSA question was easy but the
backend code was heavy to fix for the use cases, anyways couldn't complete it
Test cases passed: 5/6 for backend fix .
How was your's experience let me know
#amazonOA#connect
Got a PR merged into pytorch/torchtitan
CI was failing — torchcomms manages TP process groups outside c10d registry, so torch.compile couldn't resolve them at runtime
skipped the flavor to unblock CI until the real fix lands in torchcomms
#100DaysOfML#pytorch#opensource
Introducing Datoric: trustworthy training data for voice AI, robots, and world models.
Every collection runs in a private, project-specific workspace accessible only to invited contributors. This improves security while keeping fraud and subpar submissions out of the data pipeline. We generated almost seven figures in revenue in the last 30 days.
Introducing Datoric: trustworthy training data for voice AI, robots, and world models.
Every collection runs in a private, project-specific workspace accessible only to invited contributors. This improves security while keeping fraud and subpar submissions out of the data pipeline. We generated almost seven figures in revenue in the last 30 days.
every model run is an experiment.
today i used mlflow to track different random forest configurations, compare metrics, and save trained models instead of manually noting results.
makes it much easier to reproduce experiments later.
Just shipped CacheMind 🚀
A production-ready semantic caching middleware for LLM APIs that uses MiniLM + FAISS to retrieve cached responses through vector similarity instead of exact prompt matching.
Built with FastAPI, Redis & OpenRouter.
GitHub: https://t.co/C6Lsufrh7X
#AI#LLM
Just shipped CacheMind 🚀
A production-ready semantic caching middleware for LLM APIs that uses MiniLM + FAISS to retrieve cached responses through vector similarity instead of exact prompt matching.
Built with FastAPI, Redis & OpenRouter.
GitHub: https://t.co/C6Lsufrh7X
#AI#LLM