DSPy programmatic framework for foundation models
Framework that replaces heuristic prompt engineering with declarative modules and automated teleprompter optimizers to maximize agent pipeline accuracy.
Image: GitHub
The weekly · Week 14 · 30 Mar – 5 Apr 2026
AI coding agents: what changed
This week highlights advancements in foundation model infrastructure and observability, featuring DSPy for programmatic prompting, Meta Llama reference recipes, Xinference distributed inference, and OpenLLMetry instrumentation for AI agents.
4 picks
Meta Llama models reference implementations and tool recipes
Repository containing reference architectures, safety system prompts, and tool use recipes for the Meta Llama open model family.
Image: GitHub
Xinference distributed model inference engine
Xinference is an open-source serving system designed for running LLMs, embedding models, and multimodal architectures. It operates on CPU, GPU, and heterogeneous hardware clusters.
Image: GitHub
OpenLLMetry: OpenTelemetry Instrumentation for AI Agents
Standardized OpenTelemetry tracing extensions for monitoring agent decisions, vector queries, and API cost metrics.
Image: GitHub
Picked from public sources (GitHub, Hacker News, DEV, Lobsters, Reddit, Product Hunt) by our daily scout. Each summary uses only what the source itself says; GitHub stars are as checked on the date shown.