<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Home | Alexander Hiesch</title><description>Tinkering with the latest technologies and writing about how to apply and leverage them best.</description><link>https://hiesch.eu/</link><language>en-us</language><managingEditor>alex@hiesch.eu (Alexander Hiesch)</managingEditor><webMaster>alex@hiesch.eu (Alexander Hiesch)</webMaster><atom:link href="https://hiesch.eu/rss.xml" rel="self" type="application/rss+xml"/><item><title>CODA-DVD: Dark Vessel Detection from Orbit in Two Hours</title><link>https://hiesch.eu/blog/ai-in-space-hackathon-coda-dvd/</link><guid isPermaLink="true">https://hiesch.eu/blog/ai-in-space-hackathon-coda-dvd/</guid><description>My submission for the Liquid AI x DPhi Space &apos;AI in Space&apos; hackathon: an on-board satellite AI pipeline that filters 98% of downlink bandwidth using LFM2.5-VL.</description><pubDate>Sat, 09 May 2026 00:00:00 GMT</pubDate></item><item><title>CODA-DVD</title><link>https://hiesch.eu/projects/coda-dvd/</link><guid isPermaLink="true">https://hiesch.eu/projects/coda-dvd/</guid><description>On-board satellite AI pipeline for dark vessel detection. Three-stage cascaded filter discards clouds and empty ocean before transmission. 98% bandwidth saved.</description><pubDate>Sat, 09 May 2026 00:00:00 GMT</pubDate></item><item><title>llama.cpp on Apple Silicon: 29 GGUF Benchmarks and a 200 t/s Surprise</title><link>https://hiesch.eu/blog/llamacpp-benchmarks-speculative-decoding/</link><guid isPermaLink="true">https://hiesch.eu/blog/llamacpp-benchmarks-speculative-decoding/</guid><description>I added llama.cpp benchmarks across eight model groups to the local LLM benchmark suite. Speculative decoding with Gemma 4 hit 206 tokens per second on an M3 Max. Here&apos;s how it compares.</description><pubDate>Wed, 06 May 2026 00:00:00 GMT</pubDate></item><item><title>LLM Inference Benchmark Harness</title><link>https://hiesch.eu/projects/llm-benchmark-harness/</link><guid isPermaLink="true">https://hiesch.eu/projects/llm-benchmark-harness/</guid><description>A Python harness for systematic evaluation of LLM inference bottlenecks across backends, quantization formats, KV cache strategies, and long-context agentic workloads.</description><pubDate>Wed, 06 May 2026 00:00:00 GMT</pubDate></item><item><title>Running Local LLMs: An Easter Weekend Rabbit Hole</title><link>https://hiesch.eu/blog/running-local-llms-easter-weekend/</link><guid isPermaLink="true">https://hiesch.eu/blog/running-local-llms-easter-weekend/</guid><description>What started as a simple question (can you run agentic coding workflows against a local LLM on Apple Silicon?) turned into 57 configurations, 762 data rows, and a long weekend of benchmarking.</description><pubDate>Mon, 06 Apr 2026 00:00:00 GMT</pubDate></item><item><title>PlayAble: An Accessible Retro Gaming Portal from Hackathon Heilbronn</title><link>https://hiesch.eu/blog/hackathon-heilbronn-playable/</link><guid isPermaLink="true">https://hiesch.eu/blog/hackathon-heilbronn-playable/</guid><description>My 10-year-old and I spent 24 hours at Hackathon Heilbronn turning a 26-year-old TV game into a browser-based accessible gaming portal, without writing a single line of code ourselves.</description><pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate></item><item><title>PlayAble</title><link>https://hiesch.eu/projects/playable/</link><guid isPermaLink="true">https://hiesch.eu/projects/playable/</guid><description>An accessible retro gaming portal that lets people with disabilities play classic games through voice commands, hand gestures, or keyboard. Built in 24 hours at Hackathon Heilbronn with zero manually written code.</description><pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Three Weeks in Silicon Valley: Robotaxis, Agentic AI, and What Europe Gets Wrong</title><link>https://hiesch.eu/blog/three-weeks-silicon-valley/</link><guid isPermaLink="true">https://hiesch.eu/blog/three-weeks-silicon-valley/</guid><description>A project assignment in San Jose, friends at Applied Intuition and Zoox, and a 13-year-old at the German International School. What three weeks in Mountain View taught me about where we actually stand with AI and autonomous driving.</description><pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate></item></channel></rss>