AIAI Generated
LLM Evaluation: Test Suites That Predict Production Quality
Design production-shaped LLM evaluations with explicit tasks, calibrated graders, segmented analysis, and defensible release gates.
2026-07-16 · 14 min read
AIAI Generated
LLM Inference at Scale: Batching, KV Caches, and Serving
Engineer LLM serving around prefill and decode workloads, KV-cache memory, continuous batching, routing, autoscaling, and measurable goodput.
2026-06-07 · 14 min read
AIAI Generated
Building Reliable LLM Applications
Engineer LLM systems with typed outputs, grounded evidence, controlled tools, evals, observability, canaries, and incident-ready fallbacks.
2026-05-28 · 13 min read
AIAI Generated
Defending LLM Applications Against Prompt Injection
Defend LLM applications by modeling prompt injection as a trust-boundary and capability problem, then verify controls with adversarial tests and operational evidence.
2026-05-14 · 13 min read
AIAI Generated
Prompt Engineering as Software Engineering
Treat prompts as versioned, typed, tested production artifacts with explicit contracts, evals, observability, and security boundaries.
2025-11-17 · 11 min read
AIAI Generated
Retrieval-Augmented Generation from Scratch
Build a production-minded RAG pipeline with chunking, hybrid retrieval, reranking, citations, evaluation, security, and observability.
2025-08-28 · 14 min read
AIAI Generated
How Transformers Actually Work (No Hand-Waving)
A shape-by-shape account of tokens, attention, transformer blocks, training, inference, KV caching, and the limits behind modern LLMs.
2025-05-27 · 12 min read