Thuwarakesh MurallieinTowards Data ScienceI Tested Frontline M-LLMs on Their Chart Interpretation SkillsCan multimodal LLMs infer basic charts accurately?6d ago6d ago
Thuwarakesh MurallieinTowards Data ScienceHow Much Stress Can Your Server Handle When Self-Hosting LLMs?Do you need more GPUs or a modern GPU? How do you make infrastructure decisions?Oct 195Oct 195
Thuwarakesh MurallieinTowards Data ScienceI Fine-Tuned the Tiny Llama 3.2 1B to Replace GPT-4oIs the fine-tuning effort worth more than few-shot prompting?Oct 1527Oct 1527
Thuwarakesh MurallieinTowards Data ScienceThe Most Valuable LLM Dev Skill is Easy to Learn, But Costly to Practice.Here’s how not to waste your budget on evaluating models and systems.Oct 927Oct 927
Thuwarakesh MurallieinTowards Data ScienceBuilding RAGs Without A Retrieval Model Is a Terrible MistakeHere are my favorite techniques — one is faster, the other is more accurate.Sep 172Sep 172
Thuwarakesh MurallieinTowards Data ScienceHow I Used Clustering to Improve Chunking and Build Better RAGsIt’s both fast and cost-effectiveSep 43Sep 43
Thuwarakesh MurallieinTowards Data ScienceHow to Achieve Near Human-Level Performance in Chunking for RAGsThe costly yet powerful splitting technique for superior RAG retrievalAug 265Aug 265
Thuwarakesh MurallieinTowards Data ScienceWhy Does Position-Based Chunking Lead to Poor Performance in RAGs?How to implement semantic chunking and gain better results.Aug 22Aug 22
Thuwarakesh MurallieinTowards Data ScienceHow to Build Helpful RAGs with Query Routing.An LLM can handle general routing. Semantic search can handle private data better. Which one should you pick?Aug 162Aug 162
Thuwarakesh MurallieinTowards Data ScienceAdvanced Recursive and Follow-Up Retrieval Techniques For Better RAGsBreaking the problem solves half of it. Chaining them makes it even better.Aug 124Aug 124