
Vellum
Vellum is a YC Winter 2023 (W23) company. Based in New York, NY, USA. Team of 23.
About Vellum
Summary
Vellum builds an AI product development platform for engineering teams shipping LLM-powered applications. It covers the full build cycle: prompt management, workflow orchestration, retrieval infrastructure, automated evaluations, and production deployments, replacing a patchwork of custom scripts, spreadsheets, and manual testing. Teams use it to iterate on AI systems, measure quality and cost, and push updates to production without redeploying their application.
Go-to-market
Vellum targets B2B software teams and AI engineers, offering both a visual UI and a full-stack SDK so developers can start through self-serve access and adopt deeper tooling as their AI systems grow in complexity.
From the Y Combinator listing
The AI Product Development Platform
Prompting: Systematically iterate on prompts - Unified prompt management dashboard - Side-by-side model comparisons - Native support for tools, structured outputs, OpenAPI specs Workflows: The IDE for building steerable, agent-like systems. - Visual graph builder for orchestrating complex AI systems - Support for any model, custom code, and map/reduce functions - Built-in features like loops, parallelism, error handling Workflows SDK: A full-stack SDK + GUI for building AI apps with precision and flexibility. - Bi-directional sync between code and UI - Fine-grained control flow and global state management - Customizable with Docker and streaming support Evaluations: Test-driven development for AI. - Build scalable test suites via UI/API/CSV - Use pre-built or custom metrics for evaluation - Track quality, cost, latency, and regressions over time Retrieval: Powerful RAG infrastructure without the complexity. - Simple APIs for uploading and querying unstructured data - Tweak chunking, embeddings, and search for advanced use-cases - Supports various file types including tables and images Deployments: Update AI systems without redeploying your app. - One-click deploy across any model or provider - Staging environments for safe iteration - Scalable inference endpoints for production use Observability: Full visibility into AI system performance. - Audit logs, debugging tools, and feedback capture - Evaluation loops on live traffic - Dashboards to track cost, latency, errors, and trends
Facts
- B2B / B2B -> Engineering, Product and Design
- New York, NY, USA
- 23
- Active
- Growth
Founders (3)
- Akash Sharma
- Noa Flaherty
- Sidd Seethepalli
Score every batch against your own thesis
GPAgent's sourcing agent reads every company in a batch, scores each one against your fund's thesis in a private workspace, and lines up the founders worth meeting before Demo Day. You decide who to contact. Your scores, notes, and pipeline never appear on these pages.
How Accelerator intelligence works