Anna Alexandra Grigoryan·Jun 4, 2025Know the Ropes: Architecting Multi-Agent Systems that Scale Without ScalingThe Real Problem: When Bigger Models Stop Being SmarterA response icon2A response icon2
Anna Alexandra Grigoryan·May 28, 2025How Far Can You Push Before They Break?Instruction Saturation in LLMs
Anna Alexandra Grigoryan·May 26, 2025Multi-Label Classification with LLMs: Treat It as Classification or Don’t BotherLLMs are everywhere. But when it comes to multi-label classification, most workflows rely on hope, not rigor.A response icon1A response icon1
Anna Alexandra Grigoryan·May 10, 2025What Brain-Adapter Gets Right About Multimodal AIMultimodal AI systems are often framed as the future of deep learning. But, as we know from our work in AI, bigger isn’t always better…
Anna Alexandra Grigoryan·May 8, 2025Can LLMs Understand Graph Structure from Visuals? A Study on Layout, Prompting, and Readability
Anna Alexandra Grigoryan·May 2, 2025Doing Your Best in Layers: Adaptive Synthesis for Multi-Tier Goals in Nondeterministic Domains
Anna Alexandra Grigoryan·Apr 30, 2025Why Do Multi-Agent LLM Systems Fail?Moving the discussion around multi-agent systems from anecdote to diagnosis.
Anna Alexandra Grigoryan·Apr 30, 2025Engineering AI Risk Before It Happens: Applying Probabilistic Risk Assessment to AI SystemsAdvanced AI systems, particularly general-purpose foundation models, are moving beyond isolated applications and into core societal…
Anna Alexandra Grigoryan·Apr 18, 2025Is Asking the Right Question Harder Than Solving the Problem?Many reasoning benchmarks make a simplifying assumption: that problems are well-posed. But in production settings whether it’s in planning…
