DSPy
@DSPyOSS
Followers
12K
Following
5K
Media
87
Statuses
3K
An open-source declarative framework for building modular AI software. Programming—not prompting—LLMs via higher-level abstractions & optimizers.
Joined April 2025
DSPy is the highest-bandwidth language to talk to computers in a just-precise-enough way. First 36 lines below: 1. Take arbitrarily long content: `chunks`. 2. Build a global Table of Contents. 3. Distribute chunks into sections to be written. 4. Recursively write each section.
9
60
553
Got on a call with @mike_pavlukhin and we discussed his work using DSPy and the theme of AI in gaming https://t.co/zfml7h6XAd
0
3
8
Awesome talk from @joshclemm on how Dropbox Dash is built, including why they went with indexes rather than federated search/retrieval, using LLM as judge in evaluating retrieval quality, and using prompt optimizers like DSPy.
maven.com
Production RAG requires architectural decisions most tutorials skip: whether to index, how to structure knowledge for complex retrieval, when prompt optimization compounds value, and solving tool...
0
1
7
A customer asked this week "does Helicone work with @DSPyOSS?" So we built the integration. One line change. Full observability across your DSPy programs. Here's how 👇
1
1
7
DSPyWeekly Issue 15 - 12th Dec 2025 is out This week is absolutely stacked with: 🎙️ Omar Khattab chats with Martin Casado (a16z) on the evolution of Foundation Models @lateinteraction 📖 Early release of Mike Taylor's "Context Engineering with DSPy" book @hammer_mt 🛠️
4
10
27
@kashif_jilani Sure thing do check dspy weekly newsletter and DSPy interview series on YouTube.
2
1
8
DSPy Signature to break multi-part questions into atomic, self-contained sub-queries, plus classifies them. Post this you can decide how to handle each query. P.S. Breaking single query prompt into atomic questions is helpful and needed. The made up broad classification is
7
12
110
🚀 New technical example: @cocoindex_io + @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. @DSPyOSS provides a
9
37
258
programming, not prompting, language models
"Prompt engineering," "context engineering," "harness engineering," and whatever's next... It's one continual journey towards applying software engineering best practices (from the last half century+) to AI development.
3
4
39
🎄 Advent of Small ML: Day 9 Topic: GEPA vs. GRPO (Words vs. Weights )🎄 I've really become interested in prompt optimization methods recently because: (1) Practicality: For most real-world applications, you want to use the best possible model. Usually, that model is either
1
6
24
I just published DSPy: What ‘Programming Not Prompting’ Actually Means
0
2
12
DSPy has brought back data science into AI engineering and I love it.
5
3
88
finally getting around to testing openai's agent builder trying to define input variables. kept clicking around for how to define structured inputs. remembered, there is no such thing in openai's api. can only pass in a string cried in DSPy 😭😭😭
1
3
11