Senior Data Scientist | AI & Frontier Builder | Researcher | Blogger https://debabratamishra.github.io https://debabratamishra.github.io/posts/2026/07/graph-engineering/ https://debabratamishra.github.io/posts/2026/07/graph-engineering/ Mon, 27 Jul 2026 00:00:00 +1000 Here is the problem nobody in the multi-agent community wants to say out loud: your LLM’s context window is finite, but the world of relevant information is not. Flat, text-only context management, append tokens until you hit the limit, then... https://debabratamishra.github.io/posts/2026/04/speech-to-speech-systems/ https://debabratamishra.github.io/posts/2026/04/speech-to-speech-systems/ Sat, 11 Apr 2026 00:00:00 +1000 Over the past few years, conversational AI has been progressing at a pace that few of us could have predicted. Voice assistants no longer sound robotic, realtime translation systems can preserve the emotional nuances of a speaker’s voice, and voice... https://debabratamishra.github.io/posts/2025/09/mixture-of-experts/ https://debabratamishra.github.io/posts/2025/09/mixture-of-experts/ Sat, 20 Sep 2025 00:00:00 +1000 As Large Language Models (LLMs) continue to grow in size and complexity, a fundamental challenge emerges: how can we scale model capacity while maintaining computational efficiency? The traditional approach of simply adding more parameters to dense networks quickly becomes prohibitively... https://debabratamishra.github.io/posts/2025/07/keyvalue-caching/ https://debabratamishra.github.io/posts/2025/07/keyvalue-caching/ Fri, 25 Jul 2025 00:00:00 +1000 What is Key-Value caching? Key-value caching, as an optimization technique, focuses on improving the efficiency of the inference process in Large Language Models(LLMs) by reusing previously computed states. In simple terms, it’s a way for the model to “remember” previous... https://debabratamishra.github.io/posts/2025/07/kl-divergence/ https://debabratamishra.github.io/posts/2025/07/kl-divergence/ Sat, 19 Jul 2025 00:00:00 +1000 Understanding Kullback-Leibler divergence In today’s rapidly evolving artificial intelligence landscape, one mathematical concept stands as a cornerstone across countless applications: the Kullback-Leibler (KL) divergence. With the rise of large language models (Thanks to ChatGPT!), we’ve been fastforwarded into the future,... https://debabratamishra.github.io/posts/2025/02/blog-post-1/ https://debabratamishra.github.io/posts/2025/02/blog-post-1/ Sat, 22 Feb 2025 00:00:00 +1100 Hello World! Welcome to my first blog post. Today, I’m diving into a topic that’s shaping the future of AI — Reinforcement Learning (RL). Whether you’re an AI enthusiast, a curious reader, or just someone wondering how robots and algorithms...