# DSPy in Production DSPy is in production at **Shopify, Databricks, Dropbox, JetBlue, Moody's, AWS, Sephora, VMware**, and dozens more. This list represents companies that have publicly shared their use cases or have provided permission to be included. It reflects a selection of the many industry applications of DSPy currently in production. To add yours, please [open a PR](https://github.com/stanfordnlp/dspy/edit/main/docs/docs/community/use-cases.md). ## Featured case studies ~550× cost reduction Scaled DSPy + GEPA for structured metadata extraction across **all** Shopify shops. [Watch the talk →](https://www.youtube.com/watch?v=bxToahwOVpY) Relevance judge, automated Optimized Dash's relevance judge with DSPy for retrieval ranking and offline evaluation. [Engineering blog →](https://dropbox.tech/machine-learning/optimizing-dropbox-dash-relevance-judge-with-dspy) · [Talk →](https://www.youtube.com/watch?v=gGGCJWbqHqc) ______________________________________________________________________ ## All companies | **Company** | **Use case** | | --------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **[Shopify](https://www.shopify.com/)** | Structured metadata extraction across all Shopify shops with DSPy + GEPA; reduced yearly costs ~550×. [Talk](https://www.youtube.com/watch?v=bxToahwOVpY) | | **[Dropbox](https://www.dropbox.com/)** | Optimized Dash’s relevance judge using DSPy across ranking, training data generation, and offline evaluation. [Engineering blog](https://dropbox.tech/machine-learning/optimizing-dropbox-dash-relevance-judge-with-dspy), [Talk](https://www.youtube.com/watch?v=gGGCJWbqHqc) | | **[AWS](https://aws.amazon.com/)** | Migrates prompts from larger to smaller models on Amazon Nova while maintaining performance. [Blog](https://aws.amazon.com/blogs/machine-learning/improve-amazon-nova-migration-performance-with-data-aware-prompt-optimization/) | | **[JetBlue](https://www.jetblue.com/)** | Multiple chatbot use cases. [Blog](https://www.databricks.com/blog/optimizing-databricks-llm-pipelines-dspy) | | **[Replit](https://replit.com/)** | Synthesize diffs using code LLMs using a DSPy pipeline. [Blog](https://blog.replit.com/code-repair) | | **[Databricks](https://www.databricks.com/)** | Research, products, and customer solutions around LM judges, RAG, classification, and other applications. [Blog](https://www.databricks.com/blog/dspy-databricks) | | **[DDI](https://www.ddiworld.com/)** | Leadership coaching agents built on Databricks with DSPy. [Customer Story](https://www.databricks.com/customers/ddi) | | **[Sephora](https://www.sephora.com/)** | Undisclosed agent usecases; perspectives shared in [DAIS Session](https://www.youtube.com/watch?v=D2HurSldDkE). | | **[Zoro UK](https://www.zoro.co.uk/)** | E-commerce applications around structured shopping. [Portkey Session](https://www.youtube.com/watch?v=_vGKSc1tekE) | | **[VMware](https://www.vmware.com/)** | RAG and other prompt optimization applications. [Interview in The Register.](https://www.theregister.com/2024/02/22/prompt_engineering_ai_models/) [Business Insider.](https://www.businessinsider.com/chaptgpt-large-language-model-ai-prompt-engineering-automated-optimizer-2024-3) | | **[Nous Research](https://nousresearch.com/)** | Evolutionary self-improvement for the Hermes agent: optimizes skills, prompts, and code using DSPy + GEPA. [GitHub](https://github.com/NousResearch/hermes-agent-self-evolution) | | **[Haize Labs](https://www.haizelabs.com/)** | Automated red-teaming for LLMs. [Blog](https://blog.haizelabs.com/posts/dspy/) | | **[Plastic Labs](https://www.plasticlabs.ai/)** | R&D pipelines for Honcho. [Blog](https://blog.plasticlabs.ai/blog/User-State-is-State-of-the-Art) | | **[PingCAP](https://pingcap.com/)** | Building a knowledge graph. [Article](https://www.pingcap.com/article/building-a-graphrag-from-wikipedia-page-using-dspy-openai-and-tidb-vector-database/) | | **[Salomatic](https://langtrace.ai/blog/case-study-how-salomatic-used-langtrace-to-build-a-reliable-medical-report-generation-system)** | Enriching medical reports using DSPy. [Blog](https://langtrace.ai/blog/case-study-how-salomatic-used-langtrace-to-build-a-reliable-medical-report-generation-system) | | **[Truelaw](https://www.youtube.com/watch?v=O0F3RAWZNfM)** | How Truelaw builds bespoke LLM pipelines for law firms using DSPy. [Podcast](https://www.youtube.com/watch?v=O0F3RAWZNfM) | | **[STChealth](https://stchealth.com/)** | Using DSPy for entity resolution including human-readable rationale for decisions. | | **[Moody’s](https://www.moodys.com/)** | Leveraging DSPy to optimize RAG systems, LLM-as-a-Judge, and agentic systems for financial workflows. | | **[Normal Computing](https://www.normalcomputing.com/)** | Translate specs from chip companies from English to intermediate formal languages | | **[Procure.FYI](https://www.procure.fyi/)** | Process messy, publicly available technology spending and pricing data via DSPy. | | **[RadiantLogic](https://www.radiantlogic.com/)** | AI Data Assistant. DSPy is used for the agent that routes the query, the context extraction module, the text-to-sql conversion engine, and the table summarization module. | | **[Raia](https://raiahealth.com/)** | Using DSPy for AI-powered Personal Healthcare Agents. | | **[Hyperlint](https://hyperlint.com)** | Uses DSPy to generate technical documentation. DSPy helps to fetch relevant information and synthesize that into tutorials. | | **[Starops](https://staropshq.com/) & [Saya](https://heysaya.ai/)** | Building research documents given a user’s corpus. Generate prompts to create more articles from example articles. | | **[Tessel AI](https://tesselai.com/)** | Enhancing human-machine interaction with data use cases. | | **[Dicer.ai](https://dicer.ai/)** | Uses DSPy for marketing AI to get the most from their paid ads. | | **[Howie](https://howie.ai)** | Using DSPy to automate meeting scheduling through email. | | **[Isoform.ai](https://isoform.ai)** | Building custom integrations using DSPy. | | **[Trampoline AI](https://trampoline.ai)** | Uses DSPy to power their data-augmentation and LM pipelines. | | **[Pretrain](https://pretrain.com)** | Uses DSPy to automatically optimize AI performance towards user-defined tasks based on uploaded examples. | | **[Spindle AI](https://spindle.ai)** | Turns natural-language constrained optimization problems into solvable mathematical programs whose candidate solutions are scenarios. | | **[Infinitus](https://www.infinitus.ai/product/ai-agents/)** | Leverages DSPy to build and optimize healthcare AI agents |