Trilogy AI Center of Excellence

Driving AI innovation and practical application across Trilogy's portfolio. Explore our research, publications, and resources.

Featured Research

Highlights from our ongoing and completed research initiatives.

Generating Engaging Visuals for Education

Education

Presents a pipeline for generating adaptive educational visuals using LLM-driven multimodal models to enhance learning retention.

Multi-Agent System Coordination

Agent Systems

Explore open protocols for AI interoperability (A2A, MCP, OVON) to enable seamless collaboration across agents.

Reference-Free LLM Evaluation with G-Eval

Evaluation

Explore G-Eval metrics for reference-free evaluation using LLMs as evaluators. Examine challenges of reliability and objectivity.

Latest Publications

Stay updated with our newest findings, analyses, and insights.

[News Brief] The $100k Checkpoint, The Legacy OCR Fix, and The Antigravity Reality Check

Emerging Topics | 2025-12-05

Highlights the economics of Amazon Nova Forge, a task force win on Legacy OCR with Landing AI, and why Windsurf outclasses Google’s Antigravity alongside the launch of CoE Assist.

[Opinion] Is Nova Forge worth it?

Model Development | 2025-12-05

Spec-based critique of Amazon Nova Forge’s replay buffer and RLVR claims, questioning whether the $100k premium is genuine moat or just operational convenience.

[How-To] Why Most Architecture Review Boards Suck

Emerging Topics | 2025-12-01

Practical fixes to turn ARBs from bureaucratic bottlenecks into streamlined, AI-assisted reviews—moving from calendar-driven gatekeeping to risk-based pipelines with automation.