2026
Publications and Content, 2014–2026
The OWASP GenAI LLM Top 10 2026: What Actually Changed
Six Arm Instances, 23 Profiling Runs: A Lab Notebook
Profiling aarch64 in the Cloud: Sampling Skid, SPE, and the State of Arm Tooling
DFlash vs MTP: Two Paths to Breaking the Autoregressive Decoding Bottleneck
The Agent Productionization Gap: Why 60% of Enterprise Agentic AI Pilots Fail in 2026
Inference-Time Compute Scaling and the Overthinking Problem: How Adaptive Budget Allocation Is Reshaping Reasoning Models
EMO: How Constraining Expert Selection to Document-Level Pools Causes Semantic Modularity to Emerge During MoE Pretraining
Serverless LLM Inference Platforms: How They're Architected and What Differentiates Them
Real-World Architectures on Amazon Bedrock AgentCore
Google's A2A Protocol: Cross-Vendor Agent-to-Agent Interoperability in 2026
nanochat and the Cheap LLM Training Landscape: From $43K to $3
From RLHF to GRPO/DAPO/RLVR: How Verifiable-Reward RL Replaced Human Preference Labels
The First In-the-Wild Autonomous LLM Agent Cyberattack: What Happens When Agents Run Without Guardrails
Planner-Executor vs. Single-Loop Agents: When Does Task Decomposition Actually Help?
Speculative Decoding: How Draft-Model Parallelism Cuts LLM Inference Latency
Distillation: The Attack Vector Frontier Models Can't Hide From
vLLM Architecture & Security: A Source-Code-Level Deep Dive
Inference Runtime Security: The Hidden Attack Surface in LLM Serving
Locking Pi to Two Models: MTPLX and Ollama Cloud on Mac
Securing the AI Agent Era: DeepMind's Framework, Industry Practices & AWS Implementation
The Frontier Model Factory: How AI Is Actually Made, 2024 vs 2026
The Architecture of Intelligence: LLMs, 2024 → 2026
A Plane Radar on the Desk: The Display Was Never the Problem
The Inference Machine: KV Cache, Attention, and How Modern LLMs Stay Fast
Reading the Label: A Field Guide to Local LLM Formats, Quantization, and Variants
The Attention Tax
Running Local LLMs on a 24GB Mac: What Fits and What Doesn't
Training ML Models on AWS EKS with H100 GPUs
SSH into pihole