DSPyWeekly Issue No #15

How Foundation Models Evolved: A PhD Journey Through AI's Breakthrough Era

Omar's ( DSPy creators ) podcast episode with Martin Casado of a16z.

Context Engineering with DSPy - Book

AI engineer and educator Mike Taylor explains DSPy in a clear, approachable style, showing how its modular structure, portable programs, and built-in optimizers help teams move beyond guesswork. Through real examples and step-by-step guidance, you'll learn how DSPy's signatures, modules, datasets, and metrics work together to solve context engineering problems that evolve as models change and workloads scale.

Building Better AI Tools with MCP | Darin Kishore

Darin Kishore explores how combining Anthropic’s Model Context Protocol (MCP) with the DSPy framework can make for better AI workflow development. Kishore argues that while MCP provides a much-needed standard for connecting LLMs to external resources and tools, the real power lies in designing these tools to be "purpose-driven" and self-optimizing.

Beyond the AI Black Box by Achieving Transparency in LLM Agents with DSPy Observability - by Valdez Ladd | Dec, 2025

Adding observability to DSPy code.

Non-Obvious Things I Learned About GEPA | Elicited

Quick Primer on GEPA# GEPA (Genetic-Pareto) is a reflective optimizer. It evolves prompts by having an LLM critique failures and propose improvements. The “Pareto” in the name doesn’t mean multi-objective optimization. It means the algorithm keeps candidates that are best on any validation example, not just the best on average.

PAPER: Evolving Excellence: Automated Optimization of LLM-based Agents

We present Artemis, a no-code evolutionary optimization platform that jointly optimizes agent configurations through semantically-aware genetic operators. Given only a benchmark script and natural language goals, Artemis automatically discovers configurable components, extracts performance signals from execution logs, and evolves configurations without requiring architectural modifications. Mentions DSPy.

DSPy: Track Token Usage per-Module | Elicited

DSPy’s built-in usage tracking gives you aggregate token counts after a program runs. That’s fine for simple pipelines. But when you’re debugging cost or optimizing a multi-module agent, you need to know which predictor ate your budget.

Cocoindex plus DSPy project

New technical example: @cocoindex_io plau @DSPyOSS for structured extraction from intake forms. This demo shows how to build a fully-typed, incremental ETL pipeline where LLMs act as structured extractors — not monolithic black boxes in production systems.

Compounding Engineering (DSPy)

Compounding Engineering is a philosophy where every task you complete makes the next one easier. This isn't just about reusing code—it's about systematically extracting learnings, patterns, and decisions from each piece of work and automatically applying them to future tasks. This CLI tool implements that philosophy using DSPy

PAPER: Symbolic learning enables self-evolving agents - ScienceDirect

In this work, we introduce agent symbolic learning, a systematic framework that enables language agents to optimize themselves on their own in a data-centric way using symbolic optimizers. Mentions DSPy and uses it as baselines. Worthy read.

Early Prompt Optimization Benchmarking Results | Synth Blog

Comparing DSPy, GEPA-AI, and Synth AI on the LangProBe benchmark suite

DSPy Interview Series - Vicente on DSPy for Ruby, BAML, and the New Era of AI Engineering

In this episode, Vicente discusses his journey in developing DSPy for Ruby, the significance of BAML in AI development, and the challenges of porting libraries. He emphasizes the importance of observability in AI projects, the integration of Ruby LLM, and the evolving landscape of AI agents and workflows. Vicente also shares his vision for the future of DSPy, focusing on community engagement and education, while predicting the impact of AI on developer productivity.

Qredence/agentic-fleet

Adaptive Agentic AI Reasoning using Microsoft Agent Framework -- Join the Discord for suggestion or support ! https://discord.gg/ebgy7gtZHK | Language: Python | License: MIT License

davidberenstein1957/dspydantic

DSPydantic: Auto-Optimize Your Prompts and Pydantic Models with DSPy | Language: Python | License: Apache License 2.0

Strategic-Automation/dspy-compounding-engineering

A local-first AI engineering agent that learns from your codebase using DSPy. | Language: Python

danielckv/ContextLoom

ContextLoom is the shared "brain" for multi-agent systems. It weaves together memory threads from frameworks like DSPy and CrewAI into a unified, persistent context, powered by Redis and hydrated from your existing databases. | Language: Python

ziyacivan/braid-dspy

Python library integrating BRAID (Bounded Reasoning) into DSPy framework. | Language: Python | License: MIT License

Archelunch/dspy-toon

TOON as DSPy adapter | Language: Python

jyang297/DSPy_Adventure

Language: Python

shcallaway/dspy-gepa-example