https://astralord.github.io/A minimal, portfolio, sidebar, bootstrap Jekyll theme with responsive web design and focuses on text presentation. 2026-10-17T23:20:52+02:00 Aleksandr Samarin https://astralord.github.io/ Jekyll © 2026 Aleksandr Samarin /assets/img/favicons/favicon.ico /assets/img/favicons/favicon-96x96.png 2026-10-11T12:00:00+02:00 2026-10-17T23:20:28+02:00 https://astralord.github.io/posts/speculative-decoding-the-art-of-guessing-well/ Aleksandr Samarin

Large Language Models spend most of their decoding time not computing, but waiting: every new token requires reading all of the model weights from memory, only to multiply them by a single vector. What if the model could check several tokens for the price of one? In this post we’ll derive speculative decoding from first principles, build a general framework that covers every modern drafte...

2024-10-01T04:00:00+02:00 2026-10-11T16:37:42+02:00 https://astralord.github.io/posts/transformer-inference-optimization-toolset/ Aleksandr Samarin

Large Language Models are pushing the boundaries of artificial intelligence, but their immense size poses significant computational challenges. As these models grow, so does the need for smart optimization techniques to keep them running efficiently on modern hardware. In this post, we’ll explore key optimization strategies that are making LLMs faster and more memory-efficient. We’ll star...

2024-01-27T04:00:00+01:00 2026-05-09T11:50:35+02:00 https://astralord.github.io/posts/exploring-parallel-strategies-with-jax/ Aleksandr Samarin

Training large language models either like GPT, LlaMa or Mixtral requires immense computational resources. With model sizes ballooning into the billions or sometimes even trillions of parameters, specialized parallelization techniques are essential to make training feasible. In this post, we’ll explore implementing some of these scaling strategies in Jax - a Python framework designed for hig...

2023-07-23T18:00:00+02:00 2026-05-08T23:36:49+02:00 https://astralord.github.io/posts/building-aligned-intelligence-systems-part-ii-applying-large-language-models/ Aleksandr Samarin

In this post we will look at different techniques for steering LLMs behaviour to get desired outcomes, starting with some basic general principles such as writing a good prompt and ending with fine-tuning and augmenting models with external knowledge. Methods discussed in this post are mainly aimed at improving LLMs reliability and ensuring the consistency and factual accuracy of their...

2023-07-03T05:00:00+02:00 2026-05-09T11:50:35+02:00 https://astralord.github.io/posts/building-aligned-intelligence-systems-part-i-creating-gpt-assistant/ Aleksandr Samarin

In recent years, the field of natural language processing has witnessed a remarkable breakthrough with the advent of Large Language Models (LLMs). These models have demonstrated unprecedented performance in a wide range of language tasks, from text generation to question answering. However, the mathematical formulation of these models and the techniques used to fine-tune them for specific ta...