World's Fair 2024
Cohere for VPs of AI: Vivek Muppalla
Overview
Cohere is an enterprise AI company focused on building trustworthy AI models for real-world business use cases. The company emphasizes practical application, efficiency, and scalability over simply having the largest models. Key offerings include generative models like Command R and R Plus, and advanced retrieval models such as embeddings and rerankers. Cohere prioritizes enterprise-specific performance, customization, data privacy, and deployment flexibility across various cloud and on-premise environments.
Who should watch
- AI Engineers
- Product Managers
- Builders working on enterprise AI solutions
- Those concerned with data security and privacy in AI
- Teams looking for efficient and scalable AI models
- Developers exploring RAG and agentic use cases
Key takeaways
- Cohere offers Command R (workhorse, low-cost) and R Plus (powerful, for complex reasoning, tool use, RAG) generative models.
- Advanced retrieval models include embeddings and a specialized reranker designed to improve RAG pipeline accuracy.
- The company evaluates models based on enterprise use cases across industries like health, HR, and finance, not just academic benchmarks.
- Cohere focuses on model efficiency and scalability, aiming for practical, cost-effective deployment rather than just model size.
- Customization options range from retraining with enterprise data for domain adaptation to fine-tuning the last layers.
- Data prominence and privacy are critical, with guarantees against using customer data for model training and indemnification for IP claims.
- Deployment flexibility is a core offering, with availability on major cloud providers, on-premise, and in private VPCs.
- Features like out-of-the-box citations for RAG and strong multilingual performance (e.g., Flores evaluation) are highlighted.
Notable quotes
"We are a leading data security-focused enterprise AI company. Our focus is building trustworthy enterprise AI models with our partners for real-world business use cases."
"We're not particularly chasing the race for having the largest model out there, but what we really care about is the practical use of these models."
"Data prominence and privacy is another big focus for us."
Unofficial community note. Prefer the recording for nuance.