---
title: "Shaun Liew (@shaunliew20) | Hashnode"
description: "computer vision, vlm, agentic ai. Read the latest articles by Shaun Liew on Hashnode."
url: "https://hashnode.com/@shaunliew20"
---

@shaunliew20

# Shaun Liew

@shaunliew20 · Singapore · Joined October 2022

computer vision, vlm, agentic ai

### [Qwen3 Speculative Decoding on the DGX Spark: Two Models, Four Methods, One Surprising Lesson](https://shaunliew.hashnode.dev/qwen3-speculative-decoding-on-the-dgx-spark-two-models-four-methods-one-surprising-lesson?utm_source=hashnode&utm_medium=feed)

A beginner's guide to one of the most practical speed techniques for AI inference, tested on real hardware with real numbers.

15 min read · [shaunliew.hashnode.dev](https://shaunliew.hashnode.dev/qwen3-speculative-decoding-on-the-dgx-spark-two-models-four-methods-one-surprising-lesson?utm_source=hashnode&utm_medium=feed)

[May 18](https://hashnode.com/posts/qwen3-speculative-decoding-on-the-dgx-spark-two-models-four-methods-one-surprising-lesson/6a0a72353104e2aff0ffc755)

### [Running Qwen3-VL on DGX Spark: Transformers vs vLLM vs SGLang](https://shaunliew.hashnode.dev/running-qwen3-vl-on-dgx-spark-transformers-vs-vllm-vs-sglang?utm_source=hashnode&utm_medium=feed)

A hands-on inference benchmark across three frameworks on a desk-sized AI supercomputer.

14 min read · [shaunliew.hashnode.dev](https://shaunliew.hashnode.dev/running-qwen3-vl-on-dgx-spark-transformers-vs-vllm-vs-sglang?utm_source=hashnode&utm_medium=feed)

[May 16](https://hashnode.com/posts/running-qwen3-vl-on-dgx-spark-transformers-vs-vllm-vs-sglang/6a0824a15adec4606348738f)

### [Fixing CUDA PTX Error When Running Qwen3-VL with vLLM on H200](https://shaunliew.hashnode.dev/cuda-ptx-error-vllm-qwen3vl?utm_source=hashnode&utm_medium=feed)

Running vision-language models like Qwen3-VL with vLLM on high-end GPUs should be straightforward. Except when it's not. The Problem I was setting up Qwen3-VL-8B-Instruct on our H200 cluster (8x H200, 143GB VRAM each) when I hit this error: vllm ser...

4 min read · [shaunliew.hashnode.dev](https://shaunliew.hashnode.dev/cuda-ptx-error-vllm-qwen3vl?utm_source=hashnode&utm_medium=feed)

[Dec 7, 2025](https://hashnode.com/posts/cuda-ptx-error-vllm-qwen3vl/69354b90de71fd22f3a21015)

### [Understanding neural style transfer](https://shaunliew.hashnode.dev/understanding-neural-style-transfer?utm_source=hashnode&utm_medium=feed)

Neural Style Transfer: Turning the Mona Lisa into a Picasso Have you ever wondered what the Mona Lisa would look like if Picasso had painted it? Neural style transfer makes this artistic dream a reality by combining the content of one image with the ...

9 min read · [shaunliew.hashnode.dev](https://shaunliew.hashnode.dev/understanding-neural-style-transfer?utm_source=hashnode&utm_medium=feed)

[Jul 29, 2025](https://hashnode.com/posts/understanding-neural-style-transfer/68889c8ad16782522cd90fe0)

### [Performing an adversarial attack on images](https://shaunliew.hashnode.dev/performing-an-adversarial-attack-on-images?utm_source=hashnode&utm_medium=feed)

Exploring adversarial attacks - where tiny, imperceptible modifications can completely deceive even the smartest neural networks From Autoencoders to Adversarial Attacks: A New Kind of Magic After my fascinating journey with Variational Autoencoders,...

10 min read · [shaunliew.hashnode.dev](https://shaunliew.hashnode.dev/performing-an-adversarial-attack-on-images?utm_source=hashnode&utm_medium=feed)

[Jul 29, 2025](https://hashnode.com/posts/performing-an-adversarial-attack-on-images/68888fd5bb3a3cc57670160e)
