jagesh maharjan·Jun 7LLM Training: The 5D Parallelism UniverseA hero’s journey through the memory walls, communication storms, and clever tricks that make modern LLM training possible.A response icon1A response icon1
jagesh maharjan·Jun 7LLM Inference: The Tale of the Speed-Hungry Receipt Reader 🚀Once upon a time, I was working on a receipt extraction service powered by a 12B-parameter open-weight vision-language model. The thing…
jagesh maharjan·Jun 21, 2022Kubernetes for BeginnersEvery service is a microservice these days. And, each service stays in the containerization system (eg: Docker) as a container. The way to…A response icon3A response icon3
Indelvifybyjagesh maharjan·Jun 23, 2021How Do You Do Horizontal Partitioning in SQL?Software developers constantly optimize their infrastructure in order to gain high(er) throughput and low(er) latency…A response icon2A response icon2
Indelvifybyjagesh maharjan·May 14, 2021Let’s Deep Dive Into TransformersIn mid-2019, I published a blog on Transformers which was more of summary of the high-level architecture. As I discussed this with my…
Indelvifybyjagesh maharjan·Dec 17, 2020Large Database Scaling in MySQLDatabase scaling and elasticity is vitally important to the successful operation of any software application. A failure to scale…A response icon3A response icon3
Indelvifybyjagesh maharjan·Aug 14, 2020Implementation of Sorting Algorithms in PythonThe three most popular yet underrated algorithms in computer science are Merge Sort, Quick Sort, and Heap sort. Sounds contradiction? how…A response icon1A response icon1
Indelvifybyjagesh maharjan·Jun 9, 2020TransformerMany of our readers may have heard about BERT, the acronym for a Bidirectional Encoder Representation Transformer. Some might have used…
Indelvifybyjagesh maharjan·Mar 5, 2020YouTube Video Recommendation ArchitectureThe types of videos served by YouTube’s video recommendations have been in the news recently due to some controversial content…
Indelvifybyjagesh maharjan·Feb 26, 2020Running Mozilla’s Tensorflow Implementation of DeepSpeechIn 2014 a group at Baidu labs, we published an interesting and influential paper DeepSpeech: Scaling up end-to-end Speech RecognitionA response icon1A response icon1