Dynamic Business Logo
Home Button
Bookmark Button

AiHot100 #48 LastMile AI: AI evaluation platform

LastMile AI is a New York-based startup specializing in AI application evaluation. Founded in 2023, it offers tools for developers to test, evaluate, and benchmark AI applications. The company is recognized as one of The AI Furnace’s AI Hot 100 startups.

What does LastMile AI do?

LastMile AI provides an enterprise-grade evaluation platform designed to assist developers in testing, evaluating, and benchmarking AI applications. Its core product, AutoEval, offers out-of-the-box metrics for retrieval-augmented generation (RAG) and multi-agent AI applications, along with a fine-tuning service for custom evaluators. The platform also supports synthetic data generation, fast inference, and continuous monitoring of deployed AI models.

Key Capabilities

  • AutoEval Platform: Offers evaluation metrics for RAG and multi-agent AI applications, enabling developers to assess AI performance effectively.
  • Custom Evaluator Fine-Tuning: Allows developers to design and fine-tune evaluator models tailored to specific application criteria, enhancing evaluation accuracy.
  • Synthetic Data Generation: Automates the creation of diverse, high-quality labels, reducing manual labeling efforts and accelerating model training.
  • Real-Time Inference Infrastructure: Provides a fast inference infrastructure designed for real-time AI applications, ensuring low-latency performance.
  • Continuous Monitoring and Guardrails: Features proactive monitoring and control for deployed AI models, setting intelligent boundaries and detecting anomalies in real-time.
  • AIConfig Framework: An open-source framework for versioning, evaluating, and optimizing AI model prompts and parameters, managed as YAML configs.
  • alBERTa Language Models: Developed small language models optimized for specialized tasks, which can be fine-tuned and run efficiently on various infrastructures.

Why it stands out

LastMile AI distinguishes itself by offering a comprehensive suite of tools that address the entire lifecycle of AI application development—from evaluation and fine-tuning to deployment and monitoring. Its focus on real-time inference and continuous monitoring ensures that AI models perform reliably in production environments. Additionally, the development of alBERTa, a small language model optimized for specialized tasks, showcases the company’s commitment to innovation and efficiency in AI application development.

Quick Facts

Keep up to date with our stories on LinkedIn, Twitter , Facebook and Instagram.

What do you think?

    Be the first to comment

Add a new comment

Mazi

Mazi

Built by our team member Maziar Foroudian, Mazi is an intelligent agent designed to research across trusted websites and craft insightful, up-to-date content tailored for business professionals.

View all posts