Diagram of a pi-agent control loop with memory, tools, and a verifier.

Building a pi-agent for Long-Horizon Tasks: Patterns That Actually Work

Pi-agents for long-horizon tasks need more than a clever prompt. This post covers the architecture, memory patterns, and failure modes that decide whether the agent finishes the job or burns budget on a loop.

September 1, 2026 · 12 min · 2492 words · martinuke0
A stylized illustration of transformer blocks and a KV cache heatmap.

Build Your Own Inference Engine: From Scratch

Demystify how LLM inference works by building a tiny inference engine from scratch. Covers tokenization, the transformer forward pass, KV cache, and decoding.

September 1, 2026 · 10 min · 2116 words · martinuke0
FastAPI service diagram showing async request flow to a GPU inference worker.

FastAPI for AI Engineers: Production Patterns That Actually Ship

How to build FastAPI services that survive contact with GPU workloads, long inference calls, and real users. Patterns, anti-patterns, and production code.

September 1, 2026 · 9 min · 1811 words · martinuke0
Diagram of GMP model with goroutines, processors, and OS threads.

Inside Go's Work-Stealing Scheduler: Architecture, Goroutines, and Production Performance

How Go’s GMP scheduler assigns goroutines to OS threads, why work stealing wins for highly concurrent servers, and what to do when it doesn’t.

September 1, 2026 · 11 min · 2215 words · martinuke0
Illustration of encrypted data flowing through a network tunnel.

Mastering TLS 1.3 Zero Round-Trip Resumption: Latency Optimization, Security Trade-offs, and Real-World Implementation

Zero round‑trip resumption in TLS 1.3 can cut handshake latency dramatically while preserving strong security—learn the patterns, pitfalls, and production‑ready implementations.

June 2, 2026 · 7 min · 1359 words · martinuke0
Feedback