TinyFish has integrated its specialized web-native model development and deployment into the Databricks AI Runtime. This platform allows TinyFish to manage the entire lifecycle, from post-training and evaluation using MLflow to production serving via Databricks Model Serving. The company's model achieved a 77.6% task-success rate in internal benchmarks for its enterprise web agent. AI
IMPACT Streamlines the path from AI model development to production for enterprise web agents.
RANK_REASON Blog post detailing how a company uses a specific platform's tools for its AI workflows.
- Databricks
- Databricks AI Runtime
- Databricks Model Serving
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model
- Julia Powell
- MLflow
- Sudheesh Nair
- Tejas Sundaresan
- Tinyfish
- Zach Richardson
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →