May 6 ยท 12 min read ยท We use AI tools all the time, whether itโs asking questions, generating images, or getting help with everyday tasks. But most of these tools didnโt appear out of nowhere. They were developed based on
Join discussion
May 3 ยท 21 min read ยท TLDR: Softmax converts a vector of raw scores (logits) into a valid probability distribution by exponentiating each value and dividing by the total. Subtracting the max before exponentiating prevents floating-point overflow. Temperature scaling contr...
Join discussionMay 3 ยท 21 min read ยท TLDR: The dot product multiplies corresponding elements of two vectors and sums the results. In machine learning it does three critical jobs: it scores semantic similarity between embeddings, computes every activation in a fully connected layer, and ...
Join discussionMay 2 ยท 14 min read ยท How I combined Deep Knowledge Tracing, curriculum graph theory, and multi-objective ranking to build a recommendation system that actually understands how people learn --- and achieves a 0.0% prerequi
Join discussion
Apr 27 ยท 3 min read ยท For a long time, Iโve kept my learning and building in private. Today, Iโm changing that. I will be sharing my deep dives into ๐ง๐ฟ๐ฎ๐ป๐๐ณ๐ผ๐ฟ๐บ๐ฒ๐ฟ๐, ๐๐๐ ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ, ๐ฎ๐ป๐ฑ ๐ฅ๐๐
Join discussionApr 22 ยท 9 min read ยท March 18th, 2026. OpenAI posts the rules for something called Parameter Golf. Train the best language model that fits in a 16MB artifact and trains in under 10 minutes on 8xH100s. I read it twice and almost closed the tab. I trained a GAN a long ti...
Join discussionApr 17 ยท 26 min read ยท TLDR: Mixture of Experts (MoE) replaces the single dense Feed-Forward Network (FFN) layer in each Transformer block with N independent expert FFNs plus a learned router. Only the top-K experts activate per token โ so total parameters far exceed activ...
Join discussion
