Databricks Training

Evidence mix: High 4 · Medium 0 · Low 0 — bands are computed from each record's evidence signals.

Use cases (4)Related technologies (3)Products & vendor

Technology & SoftwareGenerative AI

Dynamo AI pretrains an 8-billion-parameter multilingual LLM in 10 days using Databricks Training

Dynamo AI, a startup building compliance-focused generative AI guardrail products, used Databricks Training to pretrain its Dynamo 8B multilingual foundation model in 10 days, achieving roughly a 20% training speedup after Databricks' AI Research team helped resolve a memory-leakage issue. The model now underpins Dynamo AI's commercial evaluation, remediation and deployment products for regulated GenAI use cases like customer support, onboarding, claims processing and fraud detection.

Dynamo AIDatabricks Training · Agent Bricks

Technology & SoftwareLarge Language ModelsConversational AIRetrieval-Augmented Generation

Democratize insight with generative AI and knowledge graphs

Stardog used Databricks Training to fine-tune large language models for Voicebox, a conversational AI interface for its Enterprise Knowledge Graph platform. After finding off-the-shelf models from OpenAI and others insufficient for the accuracy required, Stardog fine-tuned its own smaller models to match or exceed the quality of larger third-party models while retaining data security and control, letting business users query enterprise data using natural language prompts.

STStardogDatabricks Training · Agent Bricks

Technology & SoftwareCode GenerationLarge Language ModelsGenerative AI

Empowering the next billion software creators

Replit used Databricks Training infrastructure to build its own multi-billion parameter code-completion LLM from scratch, scaling up to 256 GPUs and launching the model in under a week to meet an accelerated product deadline. The model powers Replit's AI suite, which completes, explains and transforms code for its 25 million platform users, and Databricks Training reduced Replit's total cost of ownership for deploying the code-generation LLM.

REReplitDatabricks Training · Agent Bricks

Technology & SoftwareLarge Language ModelsMachine Learning

Model Experimentation Made Easier

Refuel.AI built Refuel LLM, a purpose-built model for data labeling and enrichment, by instruction-tuning a Llama-v2-13b base model on more than 5 billion tokens using Databricks Training infrastructure. The initial training run produced a 78% increase in label quality, and subsequent fine-tuning on a cluster of 8x H100s added a further 16% performance gain, outperforming trained human annotators and several other LLMs on a 15-dataset text labeling benchmark.

RERefuel.AIRefuel LLM · Llama-v2-13b · Databricks Training +1