Advanced LLM Inference Optimization Techniques
How tensor/pipeline parallelism, Grouped-Query Attention, continuous batching, and specialized serving engines overcome the GPU Memory Wall at enterprise scale.
Architecting foundation models at billion-user scale - and equipping the world's engineering teams to put AI to work. Principal ML Lead at Atlassian (ex-Twitter, Attentive, Amazon) and Founder of Afterdays - my AI consulting, technical education, and B2B AI Go-To-Market (GTM) creator practice partnering with AI Aspire (by Andrew Ng), Udacity, LinkedIn Learning, and 50+ organizations. Published author reaching millions of developers globally and recognized by the US Government (EB-1) for outstanding contributions to ML.
Closely partnering between Afterdays and Andrew Ng's AI education organization to help enterprises build practical applied ML and AI engineering skills - including delivering a 2-week corporate AI program for Morgan Stanley in New York.
Adopted in DeepLearning.AI's NLP course on Coursera, Google's TensorFlow professional certification program, and the NSF-funded StoryQ curriculum.
Comprehensive course covering model parallelism, DeepSpeed, TensorRT-LLM, and Triton Inference Server.
Designing, building, and deploying production-ready LLM-powered Recommender Systems; featured in the University of Florida Career Hub.
Enterprise AI & microservices module covering LLM chat clients, tool-calling agents (Spring AI, LangChain4j), modular RAG pipelines, and vector search (HNSW, HyDE, re-ranking).
"I was able to land an MLE FAANG offer with his reviews and prep material. 10/10 recommend!"
"He really has a range of experience and depth of knowledge that only a select few engineers can contain in one head. Could not recommend more!"
"Rishabh is an exceptionally knowledgeable mentor in the AI/ML space... Under his mentorship, I successfully completed two research papers and currently have a third paper in progress."