mchra0.bsky.social
@mchra0.bsky.social
The cuts affect nearly 10% of Amazon's 350,000 corporate staff, and making this the largest round of layoffs since 2022.

#layoffs #amazon #jobmarket
October 27, 2025 at 11:05 PM
Every post by @sebastianraschka.com goes to bookmark (secretly collecting resources)
October 16, 2025 at 5:56 PM
Do not bring in an elephant to kill an ant!
But I would love to work with more vectors and use those vector DBs and experiment. I am planning to do so shortly and post my findings.

#experiment #vectors #db #ml #ai #llm #genai
September 29, 2025 at 9:30 PM
In my case, for less than 100,000+ vectors, I can work with pgvector as it also supports HNSW and IVFFlat (although Postgres was not originally designed for vector workloads). I felt ANN performance is descent and works fine with my scenario.

#pgvector #hnsw #ivfflat #ann #vectors
September 29, 2025 at 9:30 PM
And vector DBs like Pinecone, Weaviate, Chroma & Qdrant are more suitable for high-dimensional vector similarity search and they support more optimized ANN like HNSW, IVF, PQ etc. More importantly they can work with billions of vectors with low-latency queries.

#ivf #pq #ann #hnsw #vectordb #db
September 29, 2025 at 9:30 PM
Me and my other friend had this idea to implement. But you just did it 😂 Truly amazing and thanks for showing this. This is a great learning! I’m going try it this weekend
September 18, 2025 at 4:21 PM
Can this be used for predicting next products if the products_ids are converted to semantics and passed along with context?
September 17, 2025 at 3:10 AM