<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><description>Data Science in ♥️ Home in 🇻🇳&#xA;&#xA;Kaggle Competitions Master&#xA;🥇 1 Solo Gold&#xA;🥈 2 Silvers (1 Solo, 1 Team)&#xA;🌍 Ranked 272 / 202K globally (Top 0.14%)</description><link>https://bsky.app/profile/ducnh279.bsky.social</link><title>@ducnh279.bsky.social - Duc Nguyen Huu</title><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3mfcqn437l223</link><description>When I first started learning data science, I often got lost in the @scikit-learn.org documentation. After taking the course that became this book, I understood it much better and gained confidence using it.&#xA;&#xA;Scikit-learn is a powerful library,but without guidance, it can feel overwhelming at first.&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>20 Feb 2026 18:32 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3mfcqn437l223</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3lnanglgoxs2p</link><description>&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>20 Apr 2025 12:25 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3lnanglgoxs2p</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3llzlql4kgs26</link><description>Are you familiar with Token Pooling?&#xA;&#xA;Models that use late interaction, like ColBERT, ColPali, and ColQwen, gain significant benefits from this pooling technique! By integrating token pooling methods, the number of vectors to store can be reduced.&#xA;&#xA;Blog: https://www.answer.ai/posts/colbert-pooling.html</description><pubDate>04 Apr 2025 23:41 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3llzlql4kgs26</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3llz6pv6xxk2d</link><description>Efficiently scale long CoT models like DeepSeek when using Best-of-N or Majority Voting by early pruning reasoning chains.&#xA;&#xA;Kaggle Discussion: https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-2/discussion/571669</description><pubDate>04 Apr 2025 19:48 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3llz6pv6xxk2d</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3llofcnmqv22g</link><description>I find making your agents safe is just as important as making them smart. 🔒&#xA;&#xA;A good read for building secure AI!&#xA;&#xA;arxiv.org/pdf/2503.18813</description><pubDate>31 Mar 2025 12:47 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3llofcnmqv22g</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3llmlqxredc2q</link><description>There will be one day ... in 🇺🇸 or 🇻🇳</description><pubDate>30 Mar 2025 19:37 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3llmlqxredc2q</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3ll55tamh4c25</link><description>A practical way for students to secure jobs and earn money is by developing real-world projects. Researching or engineering LLMs often seems like a field dominated by the big tech!&#xA;&#xA;It&#39;s still important to learn fundamentals from scratch for growth and problem-solving (e.g be able to fix things)! 😁&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>24 Mar 2025 16:17 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3ll55tamh4c25</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3lks7y65g2c2j</link><description>Scikit-learn accelerated 🚀&#xA;&#xA;My company has a bunch of unused T4 GPUs because the LLMs are too big for AI teams run exps. Now the data science team finally has a reason to ask for them! 🤣&#xA;&#xA;https://developer.nvidia.com/blog/nvidia-cuml-brings-zero-code-change-acceleration-to-scikit-learn/</description><pubDate>20 Mar 2025 07:57 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3lks7y65g2c2j</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3lknzrosus22s</link><description>Many good advices/best practices for missing value imputation in the paper!&#xA;&#xA;I now have a much deeper appreciation for Data School&#39;s course and regard it as the best scikit-learn course.&#xA;&#xA;Master Machine Learning with scikit-learn: https://courses.dataschool.io/master-machine-learning-with-scikit-learn&#xA;&#xA;[contains quote post or other embedded content]</description><pubDate>18 Mar 2025 15:55 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3lknzrosus22s</guid></item><item><link>https://bsky.app/profile/ducnh279.bsky.social/post/3linemyfges2t</link><description>Another great read on reasoning models!&#xA;&#xA;🧠 Small LMs struggle to learn from long or complex CoTs from larger teachers.&#xA;&#xA;🔍 Why? The reasoning complexity may be too overwhelming.&#xA;&#xA;🚀 Solution: Mix simple &amp; complex CoTs!&#xA;&#xA;📈 Results: Clear gains over complex CoTs alone!&#xA;&#xA;Arxiv: https://arxiv.org/pdf/2502.12143v1</description><pubDate>20 Feb 2025 22:46 +0000</pubDate><guid isPermaLink="false">at://did:plc:nxpvbz5j2xli4hvykf7c2jxv/app.bsky.feed.post/3linemyfges2t</guid></item></channel></rss>