<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: AndrewLiu96</title><link>https://news.ycombinator.com/user?id=AndrewLiu96</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Fri, 24 Jul 2026 03:13:16 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=AndrewLiu96" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Millwright – Rust-based, self-hosted LLM router"]]></title><description><![CDATA[
<p>No worries! Thank you for checking it out. The cache-aware concurrency was a ton of fun to implement</p>
]]></description><pubDate>Wed, 22 Jul 2026 19:12:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=49011959</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=49011959</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49011959</guid></item><item><title><![CDATA[Show HN: Millwright – Rust-based, self-hosted LLM router]]></title><description><![CDATA[
<p>Hey HN,<p>With the news of OpenRouter possibly being acquired and proliferation of hosted LLM routers (i.e. Ramp Router, Vercel’s AI Gateway), I saw the need for a self hosted solution focused on cost savings, transparency, and performance. So, I built an open sourced router with a simple CLI interface that can easily sit between coding agents and GenAI workloads.<p>For the curious and lazy, at the moment, Millwright has the tools for,<p>- Providers: OpenAI-compatible APIs, Anthropic, Amazon Bedrock<p>- Routing: policy-controlled model roles (cheap, mid, frontier), cheapest healthy route selection<p>- Protocols: OpenAI Chat Completions, Anthropic Messages, text and tool translation<p>- Cache Affinity: role-scoped session lanes without serializing concurrent agent traffic<p>- Spend Tracking: per-team costs, cache usage, model/provider mix, request traces<p>- Cost Analysis: measured usage and modeled candidate economics (HTML, Markdown, JSON)<p>- Reliability: bounded failover, circuit breakers, timeouts, concurrency limits<p>- Setup: interactive provider, model, and pricing configuration without storing provider secrets<p>- Deployment: one Rust binary, Docker, SQLite or PostgreSQL<p>Full disclosure: parts of the codebase were built with AI coding agents. All feedback is welcome, I’d especially value feedback on the routing policy, provider coverage, and anything that would block you from self-hosting it. Feel free to open feature/request and/or contribute as well.</p>
<hr>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49011806">https://news.ycombinator.com/item?id=49011806</a></p>
<p>Points: 9</p>
<p># Comments: 6</p>
]]></description><pubDate>Wed, 22 Jul 2026 19:03:50 +0000</pubDate><link>https://github.com/Northwood-Systems/millwright</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=49011806</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49011806</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>I assume you're referencing Sakana Fugu? Similar in the sense that there's multiple models involved, but from what I understand Fugu is kind of a black box in terms of what model is used for what. This project is more deterministic, you can pick and choose which model to assign/use for certain tasks, plus this project is open source which Fugu isn't.</p>
]]></description><pubDate>Thu, 09 Jul 2026 16:08:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=48848220</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48848220</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48848220</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>Thanks for checking it out!</p>
]]></description><pubDate>Thu, 09 Jul 2026 15:28:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=48847556</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48847556</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48847556</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>That's awesome! I'd love to check it out and give it a try</p>
]]></description><pubDate>Thu, 09 Jul 2026 15:28:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=48847554</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48847554</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48847554</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>Noted!</p>
]]></description><pubDate>Thu, 09 Jul 2026 15:27:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=48847546</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48847546</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48847546</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>That was actually one of my naming inspirations, the george foreman grill is a s tier invention</p>
]]></description><pubDate>Thu, 09 Jul 2026 15:26:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=48847538</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48847538</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48847538</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing"]]></title><description><![CDATA[
<p>Could you point out where you see the slop in the codebase? I'm more than happy to address it and make the tool better and more useful :)</p>
]]></description><pubDate>Wed, 08 Jul 2026 22:17:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48838128</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48838128</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48838128</guid></item><item><title><![CDATA[Show HN: Foreman, a self-hosted LLM gateway for cost aware model routing]]></title><description><![CDATA[
<p>Article URL: <a href="https://github.com/Northwood-Systems/foreman">https://github.com/Northwood-Systems/foreman</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48835063">https://news.ycombinator.com/item?id=48835063</a></p>
<p>Points: 15</p>
<p># Comments: 16</p>
]]></description><pubDate>Wed, 08 Jul 2026 17:56:53 +0000</pubDate><link>https://github.com/Northwood-Systems/foreman</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48835063</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48835063</guid></item><item><title><![CDATA[Show HN: An opinionated ranking of 21 open-weight LLMs, filterable by your GPU]]></title><description><![CDATA[
<p>Article URL: <a href="https://northwoodsystems.ai/research/open-source-models-big-board">https://northwoodsystems.ai/research/open-source-models-big-board</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48659077">https://news.ycombinator.com/item?id=48659077</a></p>
<p>Points: 3</p>
<p># Comments: 1</p>
]]></description><pubDate>Wed, 24 Jun 2026 13:01:31 +0000</pubDate><link>https://northwoodsystems.ai/research/open-source-models-big-board</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48659077</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48659077</guid></item><item><title><![CDATA[Cheaper LLM tokens led to bigger AI bills (Jevons paradox)]]></title><description><![CDATA[
<p>Article URL: <a href="https://northwoodsystems.ai/blog/ai-token-economics">https://northwoodsystems.ai/blog/ai-token-economics</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48569539">https://news.ycombinator.com/item?id=48569539</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 17 Jun 2026 12:35:19 +0000</pubDate><link>https://northwoodsystems.ai/blog/ai-token-economics</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=48569539</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48569539</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Ask HN: Who is hiring? (January 2023)"]]></title><description><![CDATA[
<p>Centro | Senior Full-Stack Engineer | Remote First (Canada) or Toronto | Canada | Full-time | <a href="https://www.centrocommerce.com/" rel="nofollow">https://www.centrocommerce.com/</a><p>Centro is hiring a Senior Full Stack Engineer to join on our engineering team. We're a venture backed startup building a next-generation data layer for e-commerce to help scaling merchants manage their inventory and orders. We are currently a small team and looking for a senior developer who wants the pace and ownership required for a fast growing startup.<p>We're looking for an experienced engineer to join our team, take on a ton of ownership & responsibly, and make a serious impact quickly. We write Python/Django on the backend and React/Javascript on the frontend. While we don't require either to apply, having experience in either will be an added bonus. We are currently only hiring candidates that are citizens or permanent residents of Canada.<p>Full job post here: <a href="https://centrocommerce.notion.site/Senior-Full-Stack-Engineer-08ad7e1aea3b4fd29f95e9d85a708177" rel="nofollow">https://centrocommerce.notion.site/Senior-Full-Stack-Enginee...</a>. Please email andrew [at] centrocommerce [dot] com with your CV to apply directly!</p>
]]></description><pubDate>Wed, 04 Jan 2023 15:25:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=34246585</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=34246585</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=34246585</guid></item><item><title><![CDATA[New comment by AndrewLiu96 in "Ask HN: Who is hiring? (December 2022)"]]></title><description><![CDATA[
<p>Centro | Senior Full-Stack Engineer | Toronto or Remote (Canada) | Canada | Full-time | <a href="https://www.centrocommerce.com/" rel="nofollow">https://www.centrocommerce.com/</a><p>Centro is hiring a Senior Full Stack Engineer to join on our engineering team. We're a venture backed startup building a next-generation data layer for e-commerce to help scaling merchants manage their inventory and orders. We are currently a small team and looking for a senior developer who wants the pace and ownership required for a fast growing startup.<p>We're looking for an experienced engineer to join our team, take on a ton of ownership & responsibly, and make a serious impact quickly. We write Python/Django on the backend and React/Javascript on the frontend. While we don't require either to apply, having experience in either will be an added bonus.<p>Full job post here: <a href="https://centrocommerce.notion.site/Full-Stack-Engineer-08ad7e1aea3b4fd29f95e9d85a708177" rel="nofollow">https://centrocommerce.notion.site/Full-Stack-Engineer-08ad7...</a>. Please email andrew [at] centrocommerce [dot] com with your CV to apply directly!</p>
]]></description><pubDate>Fri, 02 Dec 2022 22:51:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=33838062</link><dc:creator>AndrewLiu96</dc:creator><comments>https://news.ycombinator.com/item?id=33838062</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=33838062</guid></item></channel></rss>