Kelvin LuinTowards AIReally? AI Revolution is Losing Steam?Concerns of sustanability of current AI development and prediction of the future AI6 min read·1 day ago----
Kelvin LuinTowards AISora vs. Mulholland DriveAnalysis of Sora videos vs. Mulholland Drive, the human directed movie, to see how realistic the artificial generic intelligence is.10 min read·Feb 28, 2024----
Kelvin LuinTowards AIFull-Stack Data Scientist?Research of the emerging full-stack data scientists, data science engineering management, and suggestions of career development.9 min read·Feb 10, 2024--12--12
Kelvin LuinTowards AIPhi-2, the New Year Gift for Language Model LoversAn introduction to Phi-2, a compact language model that is similar to Gemini Nano. The limits of LLM, and how to fix in engineering.14 min read·Jan 21, 2024--3--3
Kelvin LuinTowards AIRegression, Personalisation, and the Kaggle SyndromeHow to do personalised prediction with XGBoost, target_encoding, recommendation, FastAI, and Kaggle dataset. It emphasises business acumen.11 min read·Nov 8, 2023--3--3
Kelvin LuinTowards AIPractical Considerations in RAG Application DesignWe discuss the practical issues when applying RAG in production, what to expect, and ways to improve RAG performance.14 min read·Oct 16, 2023--4--4
Kelvin LuinTowards AIPrompt Robustness: How to Measure and How to EnhanceDiscussion of LLM prompt robustness, measurement, enhancement, and the usage of PromptBench.10 min read·Oct 4, 2023----
Kelvin LuinTowards AIChoose Your Weapon: Survival Strategies for Depressed AI ConsultantsThis is a discussion of how the AI practitioners can cope with the rapidly advancing AI world and how to establish capability.12 min read·Sep 22, 2023--2--2
Kelvin LuDisadvantages of RAGThis is the first part of the RAG analysis:9 min read·Aug 25, 2023--9--9
Kelvin LuWhat We Need to Know Before Adopting a Vector DatabaseTo continue with our journey toward applicable Generative AI, I would like to discuss some of the challenges of applying vector databases…10 min read·Aug 15, 2023--4--4