Reference research notes on AI, infrastructure, and adjacent engineering topics. These documents are meant for orientation and quick comparison, not as final truth: AI can make mistakes, so important facts, versions, prices, and benchmarks should be checked against primary sources.
- Modern AI Models Field Guide for April 2026 - a broad guide to modern AI models: model types, architectures, naming patterns, checkpoints, quantization, runtimes, open-weight and closed families, and practical selection rules. Files: PDF, RU PDF.
- Cheat sheet for audio, speech, and music ML - an operator-oriented cheat sheet for ASR, speech emotion recognition, audio tagging, MIR, music embeddings, local inference, and hardware fit for audio/speech workloads. Files: PDF, ecosystem map PDF.
- Current State of Spiking Neural Networks 2024-2026 - an overview of SNNs: training methods, neuromorphic hardware, practical benchmarks, event-based vision, hybrid ANN-SNN approaches, and open problems. Files: PDF, RU PDF.
- Modern TypeScript Ecosystem Field Guide as of April 6, 2026 - a TypeScript/JavaScript ecosystem overview covering Node, Bun, Deno, package managers, frontend and backend frameworks, meta-frameworks, validation, testing, monorepos, and practical defaults. Files: PDF.
- Virtualization Technologies Comparison - a comparison of virtualization, containers, microVMs, and sandbox runtimes: VMware, Hyper-V, KVM/QEMU, Docker, LXC/LXD, Firecracker, gVisor, Kata Containers, WASM, and related technologies. Files: RU MD.
- Auth servers comparison - a comparison of auth and identity solutions: Keycloak, Authentik, ZITADEL, ORY, SuperTokens, Authelia, FusionAuth, Auth0, Clerk, Cognito, WorkOS, and others across self-hosting, protocols, multi-tenancy, UI, and operational complexity. Files: PDF.