A lightweight runtime health check for PyTorch training runs.

Generic modular factory takt-time simulator for process routes, buffers, transfer rules, bottlenecks, and capacity reports.

An open-source live telemetry and system modeling engine for generative AI. Track total system resource utilization in real time across storage, memory paths, and compute blocks to eliminate black-box execution failures and accelerate the development of next-generation AI and generative AI architectures.

20 Hands-on Linux performance tuning labs across physical, virtual, and cloud environments covering observability, bottleneck analysis, kernel tuning, and optimization.

Measure and visualize why LLM inference is slow: bottleneck analysis, model dissection, KV-cache, GEMM/GEMV, quantization, and memory-bound decoding.

Geomorphological investigation of the Wimbach and Klausbach valleys using ERT, TLS, GIS, geomorphic mapping, and Structure-from-Motion techniques to explore surface and subsurface landscape characteristics.

Interactive browser simulator for exploring how AI capability growth becomes real-economy output, or fails under physical and institutional bottlenecks.

Cinematic, high-performance HTML5 Canvas engine that visualizes database systems under load. Demonstrates failure, scaling, replication, caching, and optimization through real-time animation. Zero dependencies, production-grade rendering, and system architecture storytelling in a single-page app.

Cinematic, high-performance HTML5 Canvas engine that visualizes database systems under load. Demonstrates failure, scaling, replication, caching, and optimization through real-time animation. Zero dependencies, production-grade rendering, and system architecture storytelling in a single-page app.

Customer-support operations control tower: process mining, SLA bottleneck detection, and capacity forecasting on 250K+ workflow events across 50K cases. Airflow + dbt + DuckDB + Pandas + 6-page Streamlit dashboard.

Client-side GTA Online empire evaluator: income scoring, bottleneck detection, prioritized action plans, and a 795-vehicle garage database. React 19 + Vite 7, zero backend.

An advanced Python profiling and performance analysis toolkit that unifies timing, resource monitoring, line-by-line profiling, flame graphs, call graphs, memory leak detection, async/thread analysis, live dashboards, and intelligent optimization suggestions in one extensible platform.

Dependency-light OEE and bottleneck analysis for manufacturing event logs.

A.F.O artifact for bridge-sensitive bottleneck attribution and control in hierarchical LLM memory paths (gem5-based reproducibility bundle).

白毛股神的供应链瓶颈分析框架 - 5-layer supply chain bottleneck investment framework

Evidence-bound workflow diagnostics and certified lower-bound reporting for long-running AI agent pipelines. Improve agent workflow throughput without changing the model.