<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Alnur Ismail - Founder, Advisor, Investor</title>
  <subtitle>Writing by Alnur Ismail on startups, AI, and the intersection between them.</subtitle>
  <link href="https://www.alnurismail.com/"/>
  <link href="https://www.alnurismail.com/feed.xml" rel="self"/>
  <id>https://www.alnurismail.com/</id>
  <author>
    <name>ai</name>
  </author>
  <updated>2026-05-15T20:26:03.987Z</updated>

  
  <entry>
    <title>PaperClub - LLM-as-a-Judge</title>
    <link href="https://www.alnurismail.com/posts/llm-as-a-judge/"/>
    <id>https://www.alnurismail.com/posts/llm-as-a-judge/</id>
    <published>2026-05-15T00:00:00.000Z</published>
    <updated>2026-05-15T00:00:00.000Z</updated>
    <summary>Notes on LLM-as-a-judge, calibration, overconfidence, and why evaluation confidence is harder than it looks</summary>
    <category term="LLM-as-a-Judge"/><category term="Evaluation"/><category term="PaperClub"/>
  </entry>
  <entry>
    <title>We Need a New Security Paradigm for the Agentic World</title>
    <link href="https://www.alnurismail.com/posts/ai-agent-traps-into-product-tests/"/>
    <id>https://www.alnurismail.com/posts/ai-agent-traps-into-product-tests/</id>
    <published>2026-05-11T00:00:00.000Z</published>
    <updated>2026-05-11T00:00:00.000Z</updated>
    <summary>Open source framework designed to test how AI products behave when adversarial content enters their normal workflow</summary>
    <category term="Agents"/><category term="Security"/>
  </entry>
  <entry>
    <title>Poisoning AI Agents</title>
    <link href="https://www.alnurismail.com/posts/ai-agent-poison/"/>
    <id>https://www.alnurismail.com/posts/ai-agent-poison/</id>
    <published>2026-04-08T00:00:00.000Z</published>
    <updated>2026-04-08T00:00:00.000Z</updated>
    <summary>Manipulating, deceiving, and exploiting agents</summary>
    <category term="Agents"/><category term="Security"/>
  </entry>
  <entry>
    <title>PaperClub - Learning as Compilation</title>
    <link href="https://www.alnurismail.com/posts/experience-reflect-consolidate/"/>
    <id>https://www.alnurismail.com/posts/experience-reflect-consolidate/</id>
    <published>2026-03-16T00:00:00.000Z</published>
    <updated>2026-03-16T00:00:00.000Z</updated>
    <summary>Read a few papers around the reflection loop.</summary>
    <category term="Agentic"/><category term="Compilation"/><category term="PaperClub"/>
  </entry>
  <entry>
    <title>Action ⟶ Dispatch ⟶ Execute</title>
    <link href="https://www.alnurismail.com/posts/action-dispatch-execute/"/>
    <id>https://www.alnurismail.com/posts/action-dispatch-execute/</id>
    <published>2026-02-19T00:00:00.000Z</published>
    <updated>2026-02-19T00:00:00.000Z</updated>
    <summary>Realizing dispatch is not optional in tool calling systems</summary>
    <category term="Agentic"/><category term="Tools"/>
  </entry>
  <entry>
    <title>Structured Tooling for Sandboxed Agents</title>
    <link href="https://www.alnurismail.com/posts/mystro-sandboxing-tooling/"/>
    <id>https://www.alnurismail.com/posts/mystro-sandboxing-tooling/</id>
    <published>2026-02-10T00:00:00.000Z</published>
    <updated>2026-02-10T00:00:00.000Z</updated>
    <summary>Moving execution into a sandbox and rethinking tool interaction and validation</summary>
    <category term="Agentic"/><category term="Tools"/>
  </entry>
  <entry>
    <title>GEPA in the Loop</title>
    <link href="https://www.alnurismail.com/posts/mystro-gepa-run/"/>
    <id>https://www.alnurismail.com/posts/mystro-gepa-run/</id>
    <published>2026-02-05T00:00:00.000Z</published>
    <updated>2026-02-05T00:00:00.000Z</updated>
    <summary>Integrating GEPA into the agentic loop</summary>
    <category term="GEPA"/><category term="Agentic"/>
  </entry>
  <entry>
    <title>Evolutionary Prompt Optimization & Bootstrapped Coding Agents</title>
    <link href="https://www.alnurismail.com/posts/prompt-optimization/"/>
    <id>https://www.alnurismail.com/posts/prompt-optimization/</id>
    <published>2026-01-28T00:00:00.000Z</published>
    <updated>2026-01-28T00:00:00.000Z</updated>
    <summary>Building a system for evolutionary prompt optimization in agentic coding</summary>
    <category term="GEPA"/><category term="Agentic"/>
  </entry>
  <entry>
    <title>Vectors Over Tokens for Non-Invasive BCI</title>
    <link href="https://www.alnurismail.com/posts/vectors-not-tokens-for-bci/"/>
    <id>https://www.alnurismail.com/posts/vectors-not-tokens-for-bci/</id>
    <published>2026-01-16T00:00:00.000Z</published>
    <updated>2026-01-16T00:00:00.000Z</updated>
    <summary>Exploring non-invasive BCI that goes beyond backprop</summary>
    <category term="BCI"/><category term="SPA"/>
  </entry>
  <entry>
    <title>Dissecting JEPA</title>
    <link href="https://www.alnurismail.com/posts/ablation-jepa-standalone/"/>
    <id>https://www.alnurismail.com/posts/ablation-jepa-standalone/</id>
    <published>2025-12-12T00:00:00.000Z</published>
    <updated>2025-12-12T00:00:00.000Z</updated>
    <summary>Went deep down the rabit hole to validate what parts of the model work and don't</summary>
    <category term="RL"/><category term="JEPA"/>
  </entry>
  <entry>
    <title>Did JEPA Learn Anything?</title>
    <link href="https://www.alnurismail.com/posts/validating-jepa-standalone/"/>
    <id>https://www.alnurismail.com/posts/validating-jepa-standalone/</id>
    <published>2025-12-09T00:00:00.000Z</published>
    <updated>2025-12-09T00:00:00.000Z</updated>
    <summary>Trained a JEPA model and shifted focus to validating it</summary>
    <category term="RL"/><category term="JEPA"/>
  </entry>
  <entry>
    <title>JEPA Valence Standalone Experiment</title>
    <link href="https://www.alnurismail.com/posts/jepa-valence-policy-standalone/"/>
    <id>https://www.alnurismail.com/posts/jepa-valence-policy-standalone/</id>
    <published>2025-12-04T00:00:00.000Z</published>
    <updated>2025-12-04T00:00:00.000Z</updated>
    <summary>Designing an experiment to validate if JEPA can actually learn a representation</summary>
    <category term="RL"/><category term="JEPA"/>
  </entry>
  <entry>
    <title>A (attempted) JEPA Valence Policy</title>
    <link href="https://www.alnurismail.com/posts/jepa-valence-policy/"/>
    <id>https://www.alnurismail.com/posts/jepa-valence-policy/</id>
    <published>2025-12-01T00:00:00.000Z</published>
    <updated>2025-12-01T00:00:00.000Z</updated>
    <summary>Created a policy using a JEPA world model that doesn't work very well</summary>
    <category term="RL"/><category term="JEPA"/>
  </entry>
  <entry>
    <title>Paper Club - Intrinsic Motivation For RL</title>
    <link href="https://www.alnurismail.com/posts/paper-club-curiosity/"/>
    <id>https://www.alnurismail.com/posts/paper-club-curiosity/</id>
    <published>2025-11-26T00:00:00.000Z</published>
    <updated>2025-11-26T00:00:00.000Z</updated>
    <summary>Read a few papers on intrinsic motivation</summary>
    <category term="Curiosity"/><category term="PaperClub"/><category term="RL"/>
  </entry>
  <entry>
    <title>Greed Isn't a Good Policy</title>
    <link href="https://www.alnurismail.com/posts/stochastic-policy-cma-es/"/>
    <id>https://www.alnurismail.com/posts/stochastic-policy-cma-es/</id>
    <published>2025-11-24T00:00:00.000Z</published>
    <updated>2025-11-24T00:00:00.000Z</updated>
    <summary>Changing the policy to use probabilities solved the observed action space collapse</summary>
    <category term="CMA-ES"/><category term="ES"/><category term="RL"/>
  </entry>
  <entry>
    <title>Reward Shaping Bitter Lesson</title>
    <link href="https://www.alnurismail.com/posts/reward-shaping-bitter-lesson/"/>
    <id>https://www.alnurismail.com/posts/reward-shaping-bitter-lesson/</id>
    <published>2025-11-20T00:00:00.000Z</published>
    <updated>2025-11-20T00:00:00.000Z</updated>
    <summary>Introducing a more complex action space lead to bitter lesson</summary>
    <category term="CMA-ES"/><category term="ES"/><category term="RL"/>
  </entry>
  <entry>
    <title>Multilayer CMA-ES</title>
    <link href="https://www.alnurismail.com/posts/multilayer-cma/"/>
    <id>https://www.alnurismail.com/posts/multilayer-cma/</id>
    <published>2025-11-14T00:00:00.000Z</published>
    <updated>2025-11-14T00:00:00.000Z</updated>
    <summary>Understanding how to structure a multilayer CMA-ES network</summary>
    <category term="CMA-ES"/><category term="ES"/>
  </entry>
  <entry>
    <title>A Learned CMA-ES Policy to Survive</title>
    <link href="https://www.alnurismail.com/posts/cmaes-policy/"/>
    <id>https://www.alnurismail.com/posts/cmaes-policy/</id>
    <published>2025-10-31T00:00:00.000Z</published>
    <updated>2025-10-31T00:00:00.000Z</updated>
    <summary>Understanding the structure of a CMA-ES policy and training loop implementation</summary>
    <category term="CMA-ES"/><category term="RL"/><category term="ES"/>
  </entry>
  <entry>
    <title>CMA ES for Agent Training</title>
    <link href="https://www.alnurismail.com/posts/cma-es-for-agent-training/"/>
    <id>https://www.alnurismail.com/posts/cma-es-for-agent-training/</id>
    <published>2025-10-28T00:00:00.000Z</published>
    <updated>2025-10-28T00:00:00.000Z</updated>
    <summary>Applying CMA-ES to optimize agent behavior through parameter search</summary>
    <category term="CMA-ES"/><category term="RL"/><category term="ES"/>
  </entry>
  <entry>
    <title>CMA-ES Covariance Ellipse</title>
    <link href="https://www.alnurismail.com/posts/cma-es-cov-ellipse/"/>
    <id>https://www.alnurismail.com/posts/cma-es-cov-ellipse/</id>
    <published>2025-10-27T00:00:00.000Z</published>
    <updated>2025-10-27T00:00:00.000Z</updated>
    <summary>Visualizing CMA-ES search behavior and understanding step size</summary>
    <category term="CMA-ES"/><category term="ES"/>
  </entry>
  <entry>
    <title>Covariance Matrix and Distribution</title>
    <link href="https://www.alnurismail.com/posts/covariance-matrix-and-distribution/"/>
    <id>https://www.alnurismail.com/posts/covariance-matrix-and-distribution/</id>
    <published>2025-10-24T00:00:00.000Z</published>
    <updated>2025-10-24T00:00:00.000Z</updated>
    <summary>Understanding CMA-ES, NES, and multivariate Gaussian concepts</summary>
    <category term="CMA-ES"/><category term="ES"/>
  </entry>
  <entry>
    <title>Learning Without Vision</title>
    <link href="https://www.alnurismail.com/posts/learning-without-vision/"/>
    <id>https://www.alnurismail.com/posts/learning-without-vision/</id>
    <published>2025-10-22T00:00:00.000Z</published>
    <updated>2025-10-22T00:00:00.000Z</updated>
    <summary>Designing curiosity loops for simple survival-based agents</summary>
    <category term="curiosity"/><category term="motivation"/>
  </entry>
  <entry>
    <title>Evolutionary Strategies</title>
    <link href="https://www.alnurismail.com/posts/evo-strats/"/>
    <id>https://www.alnurismail.com/posts/evo-strats/</id>
    <published>2025-10-21T00:00:00.000Z</published>
    <updated>2025-10-21T00:00:00.000Z</updated>
    <summary>Exploring neuroevolution, curiosity, and the biological roots of learning</summary>
    <category term="curiosity"/><category term="evolutionary-strategies"/>
  </entry>
  <entry>
    <title>Causal World Models Overview</title>
    <link href="https://www.alnurismail.com/posts/causal-world-models-overview/"/>
    <id>https://www.alnurismail.com/posts/causal-world-models-overview/</id>
    <published>2025-09-08T00:00:00.000Z</published>
    <updated>2025-09-08T00:00:00.000Z</updated>
    <summary>Exploring causal reasoning and world models in AI</summary>
    <category term="causal-ml"/><category term="world-models"/>
  </entry>
  <entry>
    <title>Latent Space in AI</title>
    <link href="https://www.alnurismail.com/posts/latent-space-in-ai/"/>
    <id>https://www.alnurismail.com/posts/latent-space-in-ai/</id>
    <published>2025-07-23T00:00:00.000Z</published>
    <updated>2025-07-23T00:00:00.000Z</updated>
    <summary>Thinking through what latent space means and how it connects to representation learning</summary>
    <category term="latent-space"/><category term="representation-learning"/>
  </entry>
  <entry>
    <title>Text Similarity Approaches</title>
    <link href="https://www.alnurismail.com/posts/text-similarity-approaches/"/>
    <id>https://www.alnurismail.com/posts/text-similarity-approaches/</id>
    <published>2025-06-11T00:00:00.000Z</published>
    <updated>2025-06-11T00:00:00.000Z</updated>
    <summary>Exploring ASR model confidence and similarity scoring methods</summary>
    <category term="ASR"/><category term="text-similarity"/>
  </entry>
  <entry>
    <title>Understanding Latent Space</title>
    <link href="https://www.alnurismail.com/posts/understanding-latent-space/"/>
    <id>https://www.alnurismail.com/posts/understanding-latent-space/</id>
    <published>2025-06-10T00:00:00.000Z</published>
    <updated>2025-06-10T00:00:00.000Z</updated>
    <summary>Exploring methods to visualize and analyze latent spaces in VAEs and JEPA</summary>
    <category term="latent-space"/>
  </entry>
  <entry>
    <title>Encoder Neural Network Explanation</title>
    <link href="https://www.alnurismail.com/posts/encoder-neural-network-explanation/"/>
    <id>https://www.alnurismail.com/posts/encoder-neural-network-explanation/</id>
    <published>2025-02-21T00:00:00.000Z</published>
    <updated>2025-02-21T00:00:00.000Z</updated>
    <summary>Clarifying how encoders work and how to structure them for JEPA</summary>
    <category term="encoders"/><category term="JEPA"/>
  </entry>
  <entry>
    <title>Masking Strategies for JEPA</title>
    <link href="https://www.alnurismail.com/posts/masking-strategies-for-jepa/"/>
    <id>https://www.alnurismail.com/posts/masking-strategies-for-jepa/</id>
    <published>2025-02-19T00:00:00.000Z</published>
    <updated>2025-02-19T00:00:00.000Z</updated>
    <summary>Understanding deterministic masking and patching for JEPA datasets</summary>
    <category term="JEPA"/><category term="masking"/>
  </entry>
  <entry>
    <title>JEPA GTA5 World Generation</title>
    <link href="https://www.alnurismail.com/posts/jepa-gta5-world-generation/"/>
    <id>https://www.alnurismail.com/posts/jepa-gta5-world-generation/</id>
    <published>2025-02-18T00:00:00.000Z</published>
    <updated>2025-02-18T00:00:00.000Z</updated>
    <summary>Applying JEPA to generate GTA5-like worlds from frame data</summary>
    <category term="JEPA"/><category term="world-models"/>
  </entry>
  <entry>
    <title>Normalization in ML Explained</title>
    <link href="https://www.alnurismail.com/posts/normalization-in-ml-explained/"/>
    <id>https://www.alnurismail.com/posts/normalization-in-ml-explained/</id>
    <published>2025-01-31T00:00:00.000Z</published>
    <updated>2025-01-31T00:00:00.000Z</updated>
    <summary>Understanding integration, PDFs, and normalization in ML</summary>
    <category term="probability"/><category term="normalization"/>
  </entry>
  <entry>
    <title>Disentangled VAE Resources</title>
    <link href="https://www.alnurismail.com/posts/disentangled-vae-resources/"/>
    <id>https://www.alnurismail.com/posts/disentangled-vae-resources/</id>
    <published>2025-01-28T00:00:00.000Z</published>
    <updated>2025-01-28T00:00:00.000Z</updated>
    <summary>Finding research papers and resources for improving disentanglement in VAEs</summary>
    <category term="VAE"/><category term="disentanglement"/>
  </entry>
  <entry>
    <title>Dynamic VAE Layer Creation</title>
    <link href="https://www.alnurismail.com/posts/dynamic-vae-layer-creation/"/>
    <id>https://www.alnurismail.com/posts/dynamic-vae-layer-creation/</id>
    <published>2025-01-22T00:00:00.000Z</published>
    <updated>2025-01-22T00:00:00.000Z</updated>
    <summary>Building encoder and decoder layers dynamically based on input parameters</summary>
    <category term="VAE"/>
  </entry>
  <entry>
    <title>Latent Space Rotation VAE</title>
    <link href="https://www.alnurismail.com/posts/latent-space-rotation-vae/"/>
    <id>https://www.alnurismail.com/posts/latent-space-rotation-vae/</id>
    <published>2025-01-22T00:00:00.000Z</published>
    <updated>2025-01-22T00:00:00.000Z</updated>
    <summary>Exploring how to navigate or rotate within a VAE latent space</summary>
    <category term="VAE"/><category term="latent-space"/>
  </entry>
  <entry>
    <title>High Val Loss Diagnosis</title>
    <link href="https://www.alnurismail.com/posts/high-val-loss-diagnosis/"/>
    <id>https://www.alnurismail.com/posts/high-val-loss-diagnosis/</id>
    <published>2025-01-16T00:00:00.000Z</published>
    <updated>2025-01-16T00:00:00.000Z</updated>
    <summary>Understanding why validation loss is high when training a VAE and what KLD really means</summary>
    <category term="VAE"/><category term="KLD"/><category term="loss-analysis"/>
  </entry>
  <entry>
    <title>VAE RGB Importance</title>
    <link href="https://www.alnurismail.com/posts/vae-rgb-importance/"/>
    <id>https://www.alnurismail.com/posts/vae-rgb-importance/</id>
    <published>2025-01-16T00:00:00.000Z</published>
    <updated>2025-01-16T00:00:00.000Z</updated>
    <summary>Debugging color and visualization issues when training VAEs on RGB data</summary>
    <category term="VAE"/><category term="CV"/>
  </entry>
  <entry>
    <title>Learning VAEs and World Gen</title>
    <link href="https://www.alnurismail.com/posts/learning-vaes-and-projects/"/>
    <id>https://www.alnurismail.com/posts/learning-vaes-and-projects/</id>
    <published>2025-01-14T00:00:00.000Z</published>
    <updated>2025-01-14T00:00:00.000Z</updated>
    <summary>Starting to learn VAEs and figuring out how to apply them to world generation</summary>
    <category term="VAE"/><category term="world-models"/>
  </entry>
  <entry>
    <title>Approximate Inference Simplified</title>
    <link href="https://www.alnurismail.com/posts/approximate-inference-simplified/"/>
    <id>https://www.alnurismail.com/posts/approximate-inference-simplified/</id>
    <published>2025-01-07T00:00:00.000Z</published>
    <updated>2025-01-07T00:00:00.000Z</updated>
    <summary>Breaking down probabilistic models and VAEs to understand inference, ELBO, and KL divergence</summary>
    <category term="VAE"/><category term="KLD"/>
  </entry>
  <entry>
    <title>GQA vs MQA</title>
    <link href="https://www.alnurismail.com/posts/learning-about-emergent/"/>
    <id>https://www.alnurismail.com/posts/learning-about-emergent/</id>
    <published>2025-01-02T00:00:00.000Z</published>
    <updated>2025-01-02T00:00:00.000Z</updated>
    <summary>Understanding exponential forms and attention variants in transformers</summary>
    <category term="attention"/><category term="transformers"/>
  </entry>
  <entry>
    <title>Probabilistic Programming</title>
    <link href="https://www.alnurismail.com/posts/probabilistic-programming-learning-plan/"/>
    <id>https://www.alnurismail.com/posts/probabilistic-programming-learning-plan/</id>
    <published>2024-12-09T00:00:00.000Z</published>
    <updated>2024-12-09T00:00:00.000Z</updated>
    <summary>Building a learning plan for probabilistic programming in AI</summary>
    <category term="probabilistic-programming"/><category term="research"/>
  </entry>
  <entry>
    <title>Novel AI Approaches</title>
    <link href="https://www.alnurismail.com/posts/novel-ai-approaches/"/>
    <id>https://www.alnurismail.com/posts/novel-ai-approaches/</id>
    <published>2024-11-26T00:00:00.000Z</published>
    <updated>2024-11-26T00:00:00.000Z</updated>
    <summary>Looking at new directions in AI that go beyond deep neural networks</summary>
    <category term="research"/>
  </entry>
  <entry>
    <title>Understanding State-Space Models</title>
    <link href="https://www.alnurismail.com/posts/understanding-state-space-models/"/>
    <id>https://www.alnurismail.com/posts/understanding-state-space-models/</id>
    <published>2024-10-29T00:00:00.000Z</published>
    <updated>2024-10-29T00:00:00.000Z</updated>
    <summary>Breaking down Mamba’s state equations and how they relate to ML model outputs</summary>
    <category term="mamba"/><category term="SSM"/>
  </entry>
  <entry>
    <title>Wav2Vec2 Has 12 Layers</title>
    <link href="https://www.alnurismail.com/posts/wav2vec2-has-12-layers/"/>
    <id>https://www.alnurismail.com/posts/wav2vec2-has-12-layers/</id>
    <published>2024-07-04T00:00:00.000Z</published>
    <updated>2024-07-04T00:00:00.000Z</updated>
    <summary>Inspecting transformer layers and hidden states in Wav2Vec2</summary>
    <category term="ASR"/><category term="wav2vec"/>
  </entry>
  <entry>
    <title>More Epochs Better Inference</title>
    <link href="https://www.alnurismail.com/posts/more-epochs-better-inference/"/>
    <id>https://www.alnurismail.com/posts/more-epochs-better-inference/</id>
    <published>2024-06-19T00:00:00.000Z</published>
    <updated>2024-06-19T00:00:00.000Z</updated>
    <summary>Understanding the tradeoffs between epochs, data, and overfitting</summary>
    <category term="training"/><category term="generalization"/>
  </entry>
  <entry>
    <title>Agentic Thinking</title>
    <link href="https://www.alnurismail.com/posts/agents-thinking/"/>
    <id>https://www.alnurismail.com/posts/agents-thinking/</id>
    <published>2024-06-13T00:00:00.000Z</published>
    <updated>2024-06-13T00:00:00.000Z</updated>
    <summary>Connecting agentic workflows to tree-of-thought and reinforcement learning ideas</summary>
    <category term="agentic-systems"/><category term="reasoning"/>
  </entry>
  <entry>
    <title>SpeechBrain Architecture</title>
    <link href="https://www.alnurismail.com/posts/speechbrain-structure/"/>
    <id>https://www.alnurismail.com/posts/speechbrain-structure/</id>
    <published>2024-06-13T00:00:00.000Z</published>
    <updated>2024-06-13T00:00:00.000Z</updated>
    <summary>Exploring SpeechBrain's modular system of lobes and recipes for ASR and representation learning.</summary>
    <category term="speechbrain"/><category term="ASR"/>
  </entry>
  <entry>
    <title>Audio Feature Interpretability</title>
    <link href="https://www.alnurismail.com/posts/audio-feature-inter/"/>
    <id>https://www.alnurismail.com/posts/audio-feature-inter/</id>
    <published>2024-05-16T00:00:00.000Z</published>
    <updated>2024-05-16T00:00:00.000Z</updated>
    <summary>Investigating how to reconstruct or understand the 768-dim latent features from wav2vec 2.0.</summary>
    <category term="ASR"/><category term="latent-space"/><category term="wav2vec"/>
  </entry>
  <entry>
    <title>Representation Learning in wav2vec Models</title>
    <link href="https://www.alnurismail.com/posts/rep-learning-in-wav2vec/"/>
    <id>https://www.alnurismail.com/posts/rep-learning-in-wav2vec/</id>
    <published>2024-05-16T00:00:00.000Z</published>
    <updated>2024-05-16T00:00:00.000Z</updated>
    <summary>Exploring how wav2vec 2.0 and vq‑wav2vec learn speech representations via self‑supervised objectives.</summary>
    <category term="wav2vec"/><category term="ASR"/>
  </entry>
  <entry>
    <title>Phoneme-level ASR models</title>
    <link href="https://www.alnurismail.com/posts/phoneme-models/"/>
    <id>https://www.alnurismail.com/posts/phoneme-models/</id>
    <published>2024-05-02T00:00:00.000Z</published>
    <updated>2024-05-02T00:00:00.000Z</updated>
    <summary>Evaluating SpeechBrain's support for phoneme-level ASR and model discovery challenges.</summary>
    <category term="ASR"/><category term="phoneme-recognition"/>
  </entry>
</feed>