Blog
Technical exploration and thoughts · 438 articles
Hive Slowly Changing Dimension Type 2: Order History State Incremental Refresh and Close Chain Logic
Offline data warehouse needs to save order history state at low cost while supporting daily rollback and change analysis. This article introduces using ODS...
Rebuilding My Personal Site with Astro: Tech Stack Choices and Pitfalls
Documenting the complete process of migrating wzk.icu from plain HTML to Astro 6: why I chose Astro, content collection setup, Bun gotchas, and SEO infrastructure building.
LLM Application Engineering: Key Practices from Demo to Production
Core experience moving LLM applications from prototype to production: context management, error handling, cost control, observability. No basics, just real pitfalls.
Real-time Voice Interaction Pipeline Latency Optimization Practice
Documenting the process of building ASR→LLM→TTS real-time voice pipeline: why latency is high, how pipeline concurrency reduces first-byte latency, VAD endpoint detection pitfalls, and practical component selection experience.
Deploying Spring Boot on K8s: 5 Pitfalls I Encountered
Documenting real issues encountered when migrating Spring Boot microservices to Kubernetes: health checks, ConfigMap hot-reload, JVM memory, graceful shutdown, and log collection.
Spring Cloud Gateway Rate Limiting and Circuit Breaker Practice
Recording the complete process of configuring Spring Cloud Gateway rate limiting (Redis token bucket) and circuit breaker (Resilience4j) in production, including pitfalls and tuning experience.
ESP32 Integration with Home Assistant: Building Temperature and Humidity Monitoring from Scratch
Build a temperature and humidity sensor using ESP32 + DHT22, integrate with Home Assistant via MQTT, and achieve automation. Complete steps including troubleshooting guide.
AI Research #135: Gemini 3 Pro Back on Top - MoE, Million Context
Explains Gemini 3 Pro's advantages through sparse MoE architecture, million-token context, native multimodal (text/image/video/PDF), thinking depth control (thinking_level), and Deep Think mode. Stable advantage in high-difficulty reasoning and long-horizon tasks.
AI Research #134: Java 2025 - Will It Decline?
In 2025, Java remains the enterprise backend and critical industry workhorse. Key keywords: Java 25 LTS, Java 21 LTS, Spring Boot, MicroProfile, Kubernetes, Serverless, Project Loom, GraalVM Native Image, CRaC, Project Leyden, Panama.
AI Research #133: Java vs Kotlin/Go/Rust/Python/Node 2025
Comprehensive comparison of Java 21/23 (Virtual Threads/Loom) with Kotlin 2.x, Go 1.22/1.23, Rust 1.7x, Python (PyTorch 2.x/TensorFlow 2.x), Node.js 20/22. Covers ecosystem, performance, and application scenarios. Java maintains best overall value in enterprise backend and high-concurrency services.
AI Research #132: Java Ecosystem 2025 - Spring, Quarkus, GraalVM, CRaC
Spring Framework 6, Spring Boot 3.x (minimum JDK 17, supports Java 21 Virtual Threads), GraalVM Native Image AOT, Quarkus and Micronaut cloud-native...
AI Research #131: Java 17/21/25 Complete Comparison
Java 17 (2021), Java 21 (2023), Java 25 (2025) language and JVM changes, covering Virtual Threads (Project Loom), Records/Pattern Matching (Project Amber),...
AI Research #130: Qwen2.5-Omni Practical Applications
Office assistant, education and training, programming and operations, search-enhanced RAG, device control/plugin agents, and companion entertainment. Covers...
AI Research #129: Qwen2.5-Omni-7B Key Specs - VRAM, Context, Concurrency
Runs stably at FP16 ~14GB VRAM, with INT8/INT4 quantization (<4GB) enabling deployment on consumer GPUs or edge devices. Combined with FlashAttention 2 and...
AI Research #128: Qwen2.5-Omni Training Pipeline - Three-Stage Alignment
Complete training pipeline breakdown for Qwen2.5-Omni: Thinker based on Qwen2.5, vision initialized from Qwen2.5-VL, audio from Whisper-large-v3. Uses...
AI Research #127: Qwen2.5-Omni Deep Dive - Thinker-Talker Architecture
Engineering breakdown of Qwen2.5-Omni (2024-2025) Thinker-Talker dual-core architecture: unified Transformer decoder for text/image/video/audio fusion, TMRoPE...
AI Research #126: Tesla FSD V14 Engineering - High-Precision Vision Occupancy Network
Tesla FSD V14 related US patents: engineering analysis of high-precision vision occupancy network (SDF ~10cm), sparse sampling + trilinear interpolation,...
AI Research #125: Tesla FSD Business Model and Competitor Comparison
FSD V14 (2025) business model and competitive landscape. Analyzes pricing logic for one-time purchase (~15,000) vs subscription (~199/month) and deferred...
AI Research #124: Tesla FSD V14 Deep Analysis
Tesla FSD V14 real-world performance and road tests, comparing V13.2 on urban roads and highways: key disengagement metrics, lane changes/ramps, destination arrival, and long-tail scenarios. V14 shows significant improvement in miles per disengagement on both urban and highway.
AI Research #123: FSD V14 Deep Analysis - Vision-Only SDF vs V12 Upgrades
FSD V14 (2025) technical evolution compared to V12 (2023), focusing on vision-only approach, SDF (Signed Distance Field) occupancy reconstruction, end-to-end...