---
title: "Anton R Gordon (@AntonRGordon) | Hashnode"
description: "Data Architect | Cloud Specialist | Technical Leader. Read the latest articles by Anton R Gordon on Hashnode."
url: "https://hashnode.com/@AntonRGordon"
---

@AntonRGordon

# Anton R Gordon

@AntonRGordon · Houston, Texas · Joined May 2024

Data Architect | Cloud Specialist | Technical Leader

[Website](https://antongordon.com/)

### [From GPU Utilization to GPU Economics: What We Learned Building a Distributed AI Performance Assessment](https://antonrgordon.hashnode.dev/from-gpu-utilization-to-gpu-economics-what-we-learned-building-a-distributed-ai-performance-assessment?utm_source=hashnode&utm_medium=feed)

Why measuring tokens per second, NCCL behavior, and communication exposure together tells a much more useful story than GPU utilization alone. The easiest way to misunderstand distributed AI performan

9 min read · [antonrgordon.hashnode.dev](https://antonrgordon.hashnode.dev/from-gpu-utilization-to-gpu-economics-what-we-learned-building-a-distributed-ai-performance-assessment?utm_source=hashnode&utm_medium=feed)

[Oct 10](https://hashnode.com/posts/from-gpu-utilization-to-gpu-economics-what-we-learned-building-a-distributed-ai-performance-assessment/6aca31a65407e0f47c81da08)

### [Anton R Gordon on Prefill vs. Decode: Finding the Real Bottleneck in LLM Inference](https://antonrgordon.hashnode.dev/anton-r-gordon-on-prefill-vs-decode-finding-the-real-bottleneck-in-llm-inference?utm_source=hashnode&utm_medium=feed)

By Anton R Gordon When an LLM endpoint feels slow, the instinct is usually to blame the model. The model is too large. The GPU is too slow. The context window is too long. We need more GPUs. But those

12 min read · [antonrgordon.hashnode.dev](https://antonrgordon.hashnode.dev/anton-r-gordon-on-prefill-vs-decode-finding-the-real-bottleneck-in-llm-inference?utm_source=hashnode&utm_medium=feed)

[Sep 27](https://hashnode.com/posts/anton-r-gordon-on-prefill-vs-decode-finding-the-real-bottleneck-in-llm-inference/6ab91e1976747f4c7c1017c7)

### [Anton R Gordon on Why Modern LLMs Think in Context, Not Sequences: The Evolution Beyond Self-Attention](https://antonrgordon.hashnode.dev/anton-r-gordon-on-why-modern-llms-think-in-context-not-sequences-the-evolution-beyond-self-attention?utm_source=hashnode&utm_medium=feed)

When the Transformer architecture was introduced in the landmark 2017 paper Attention Is All You Need, it fundamentally changed how machines process language. Instead of reading text one token at a ti

5 min read · [antonrgordon.hashnode.dev](https://antonrgordon.hashnode.dev/anton-r-gordon-on-why-modern-llms-think-in-context-not-sequences-the-evolution-beyond-self-attention?utm_source=hashnode&utm_medium=feed)

[Jul 4](https://hashnode.com/posts/anton-r-gordon-on-why-modern-llms-think-in-context-not-sequences-the-evolution-beyond-self-attention/6a4875908eda97b8bc7cfd53)

### [Anton R Gordon on Building AI-Powered Equity Research Systems That Analysts Actually Trust](https://antonrgordon.hashnode.dev/anton-r-gordon-on-building-ai-powered-equity-research-systems-that-analysts-actually-trust?utm_source=hashnode&utm_medium=feed)

Artificial intelligence has become a common fixture in financial services, but most AI-powered research tools suffer from the same fundamental problem: they generate answers that sound convincing with

4 min read · [antonrgordon.hashnode.dev](https://antonrgordon.hashnode.dev/anton-r-gordon-on-building-ai-powered-equity-research-systems-that-analysts-actually-trust?utm_source=hashnode&utm_medium=feed)

[Jun 20](https://hashnode.com/posts/anton-r-gordon-on-building-ai-powered-equity-research-systems-that-analysts-actually-trust/6a368d7fa6de6eaf3d12eb63)

### [How Anton R Gordon Designs Low-Latency AI Systems for High-Frequency Decision Environments](https://antonrgordon.hashnode.dev/how-anton-r-gordon-designs-low-latency-ai-systems-for-high-frequency-decision-environments?utm_source=hashnode&utm_medium=feed)

In modern enterprise environments, AI systems are no longer operating in isolated analytics workflows. They are increasingly embedded into environments where decisions must happen in milliseconds—fina

4 min read · [antonrgordon.hashnode.dev](https://antonrgordon.hashnode.dev/how-anton-r-gordon-designs-low-latency-ai-systems-for-high-frequency-decision-environments?utm_source=hashnode&utm_medium=feed)

[May 11](https://hashnode.com/posts/how-anton-r-gordon-designs-low-latency-ai-systems-for-high-frequency-decision-environments/6a01ad0cfca21b0d4b0015b4)