AI Engineering

A method for agents to persist critical state—facts, failures, conclusions—in a file that survives context compaction and session ends, with strategies for retrieval and preventing stale data.

Manage what an LLM sees by budgeting the context window, retrieving relevant information, compacting history, and pruning tool results to prevent attention degradation.

Guides the decision to fine-tune versus prompt engineering or retrieval, and provides a rigorous process for dataset construction, training with LoRA or full fine-tuning, and evaluation to achieve reliable performance improvements.

Enforce rules deterministically in AI agents using lifecycle hooks. Automate formatting, block dangerous commands, and inject context without depending on model memory.

Optimize LLM feature costs by identifying where tokens are spent and applying caching, context reduction, model routing, batching, and output limits without degrading output quality.

Provides a systematic approach for measuring LLM feature quality, including building evaluation sets, selecting metrics, using LLM-as-judge, regression testing, and production evaluation. Use before iterating on prompts or models to avoid guesswork.

Use when integrating an LLM API into an application. Covers streaming, retries and rate limits, timeouts, caching, fallback across providers, and the production concerns that a tutorial integration ignores.

Use when building a Model Context Protocol server. Covers tool, resource, and prompt design, transport choice, authentication, error handling, and testing against a real client.

Helps select the right LLM for a task by evaluating capability, cost, latency, and routing strategies. Use when choosing a model for a new feature, reducing costs, or evaluating model migrations.

Helps researchers check if their idea has been scooped and browse daily new papers via arXiv and user-submitted sources. Presents objective concept overlap analysis, leaving decisions to the user.

Helps researchers find academic papers, identify research gaps, and track citations in both English and Chinese, using automated API retrieval for English databases and guided manual retrieval for Chinese databases.

Write and improve prompts for language models. Covers instruction design, few-shot examples, reasoning elicitation, output formatting, and systematic failure diagnosis.