Localization Today Rewriting the MT Playbook with Marco Trombetti
May 12, 2025
Marco Trombetti, CEO and Co-Founder of Translated, shares insights on LARA, their cutting-edge translation system, which boosts accuracy and fluency in 32 languages. He discusses how AI agents are set to revolutionize localization, not just through translation but also by streamlining project management. The conversation dives into the MCP protocol's role in fostering collaboration among tech giants and highlights how new features in LARA can enhance team efficiency. Marco emphasizes the need for open standards to tackle interoperability challenges in the digital agent landscape.
AI Snips
Chapters
Transcript
Episode notes
LARA Blends LLM and Translation Models
- Translated's LARA combines large language models (LLMs) with translation models to merge fluency with precision.
- This synergy eliminates the tradeoff between fluency and accuracy common in machine translation systems.
LARA's Advanced Language Features
- LARA now supports 32 languages with advanced capabilities like adaptation, context use, and instructions.
- Users can adapt models in real-time with translation memory integration, improving flexibility and translation quality.
Centralize Translation Control with Team Plan
- Use LARA's Team Plan to centralize terminology, style, and model adaptation across all company users.
- This empowers localization teams to control translation quality and extend MT benefits organization-wide.