Latest News and Insights

Exploring Foundation Models' Tool-Use Efficacy

Model Context Protocol (MCP) is an open protocol launched by Anthropic to standardize the way LLMs use external tools. AI agents use MCP to enable multi-turn workflows, where an LLM (via products like [Claude Desktop](https://claude.ai/download) or [Cursor](https://docs.cursor.com/en/tools/mcp)) can select and coordinate between tools in multiple MCP servers. Since its introduction, MCP has quickly become the de facto standard for tool integrations with LLMs.

Applying RL: Improving Code Merging

While foundation models have continued to improve at coding capabilities, using foundation models for high specificity, low complexity tasks like code merging can be overkill. With that in mind, we saw an opportunity to use reinforcement learning to fine-tune a model (Qwen3-1.7B) for code merge - the result is a small model that’s better and faster than foundation models, while also able to run locally.

Applying RL: Fixing Structured Outputs

A significant portion of AI use cases revolve around structured outputs - i.e. using the model to ingest unstructured textual data to generate a structured output, typically in JSON format. However, this leads to a performance decrease in tasks that are not strictly just formatting changes since structured output mode enforces a schema and stops the model from thinking ‘freely’.

Applying RL: Open Source SLM Trained for MCP

MCP is quickly becoming the open standard for AI agents, and for good reason! It’s well designed and easy to use — however, MCP implementations are limited since: 1. The best models are large and closed-source (3.7 Sonnet, Gemini 2.5 Pro) 2. It introduces tool sprawl (numerous MCP clients, MCP servers to integrate)

Responsive Software

Software has changed a lot since its infancy in the 1950s. However – from early mainframes and monolithic architectures to modern unstructured databases and distributed architectures – the fundamental pattern has remained the same: moving data to a central location for processing. The technology has changed from mainframes to SQL databases to cloud data warehouses & lakehouses, but the core pattern of data centralization persists.

Teach AI Agents How to Learn in Real-Time