Knowledge Base
Original articles about everyday developer problems: the principles, the practical steps, and the tools that pair with them.
Browse by topic
Using LLMs in real development: prompt structure, output verification, local vs cloud trade-offs, and the limits of RAG and context windows.
- What Is Jev: An AI Decision Model That Does Not Write
- Local LLM vs Cloud API: Privacy, Cost and Capability Trade-offs
- RAG and Context Windows: Show the Model Only What It Needs
Encoding vs encryption, hashing and password storage, symmetric and asymmetric crypto, HTTPS and JWT in practice.
- How to Verify File Integrity with SHA-256
- Strong Passwords: Why Length Beats Complexity
- Base64 Is Not Encryption: Encoding vs Encryption
Base64, URL encoding, JSON, timestamps, number bases and colour formats — specs, conversion and pitfalls.
- Number Bases Made Intuitive: Binary, Hex and Color Codes
- Unix Timestamps 101: Seconds vs Milliseconds, UTC and Timezones
- Character Encoding and UTF-8: Why Text Turns into Gibberish
IP geolocation and route analysis, Cron scheduling, and practical approaches to network troubleshooting.
- API Idempotency, Retries and Timeouts: Making Failure Predictable
- IP Route Analysis: What CN2 / 163 / 9929 / CMI Mean
- HTTP Status Codes and Caching Headers: A Practical Guide
Local-first privacy architecture, OCR accuracy, watermarking and identifier design — hands-on engineering notes.
- Everyday Git and Recovering from Four Common Accidents
- Add a Watermark Before You Share Images
- Practical Regex Recipes and Four Common Traps
Latest articles
Jev generates no text: it turns decisions into structured values software can use directly. What it is, why it appeared, and when not to use it.
There is no universally best option, only a scenario-matched one. Compare privacy, capability, cost and ops — plus a practical hybrid path.
A model does not know your internal docs or recent events. The minimal RAG pipeline, context-window trade-offs, and five common pitfalls.
Same prompt, different result? The requirement was never constrained. Five prompt modules, few-shot examples, and hard output rules.
Hallucination cannot be prompted away. Four high-risk scenarios and a six-step checklist: models propose, tools and humans confirm.