Avanish Garg is an Applied AI Engineer with over 2 years of experience in building and deploying production AI/ML systems, specializing in LLM fine-tuning, retrieval-augmented generation (RAG), and agentic pipelines. He has a proven track record of optimizing model quality, latency, and inference costs at scale, supported by significant contributions to major AI/ML open-source projects. Currently, he is focused on transforming research techniques into reliable, reusable AI infrastructure, while also engaging in technical writing to demystify complex systems in AI and Web3.
- AI
- Web3
- machine learning
- building AI infrastructure
- optimizing model performance
- developing AI services
- AI researchers
- open-source contributors
- developers in AI
“Your AI does not remember you. I just published Ep 05 of The Curiosity Crunch, deconstructing the architecture of context windows, self-attention, and why faking AI memory is computationally expensive.”
“AI does not understand a single word you type. It is a collective hallucination by high-dim calculus and floating-point arrays. I just published Ep 04 of The Curiosity Crunch, deconstructing the vector math and latent space behind the illusion.”