EXPERT COURSE

AI Translation:
NMT, LLMs, MTQE+APE

AI Translation: Fastest Transformation in History

The AI-driven translation landscape is undergoing the fastest transformation in its history.
Neural Machine Translation (NMT) has become standard across the industry, Large Language Models (LLMs) are introducing new paradigms of multilingual reasoning, and hybrid workflows such as MTQE and Automated Post-Editing (APE) are redefining translation quality management.
As organizations search for scalability, speed, and cost efficiency, professionals who understand how to integrate and evaluate these technologies are becoming indispensable.

85% of Localization Teams Deploying AI for Translation

According to CSA Research reports, 85% of enterprise localization teams are already deploying some form of machine translation, while Gartner notes that LLM-based automation will reshape 40% of localization workflows by 2027. Meanwhile, according to Slator, more than 70% of LSPs consider hybrid MT+LLM workflows as their primary investment area for the coming years. These shifts are redefining

Become a Pro in AI for Translation

This comprehensive 14-hour course equips you with the knowledge and applied skills required to navigate this new ecosystem confidently.

You will learn:
  • how NMT engines are trained and fine-tuned
  • how adaptive MT and terminology injection function in real-world workflows
  • how evaluation frameworks (both human and automated) integrate into scalable quality strategies
  • how LLM prompting, RAG, and fine-tuning can be applied specifically to translation tasks.


Beyond theory, you will explore the advanced MTQE–APE hybrid workflow, now considered the cornerstone of high-throughput, low-touch production pipelines in modern localization operations.

You will also complete a practical hands-on session using OpenAI Agent Builder to simulate MTQE and APE automation logic.


Conditions: Please read our course and subscription plans terms and conditions carefully. With your registration, you confirm that you have read, understood and accepted our conditions and agree with them. 

If you have any questions, please visit the FAQ section (for courses or subscription plans) or get in touch with us.

Add to Calendar

Apple Google Office 365 Outlook Outlook.com Yahoo

  • This course includes:
  • Expert coach:
    Andrés Romero Arcas, Language Technology Expert at Acolad
  •  Interactive activities
  •  Life access to contents 
  •   Downloadable course program
  •  In English
  •  Completion certificate
  •  Money back guarantee
  • Acquire A-Z knowledge about AI Translation
  • When: 2-10 February 2026 - 18.00 CET
  • Duration: 12 h approx.

Course and Instructor Description

The AI Translation (NMT, LLMs, MTQE+APE) course provides a complete introduction to today’s most advanced translation technologies.
Participants will learn:
  • how Neural Machine Translation engines are built and fine-tuned
  • how Large Language Models can be applied to translation tasks
  • how hybrid workflows, such as Machine Translation Quality Estimation (MTQE) and Automatic Post-Editing (APE), shape scalable, high-quality localization pipelines.

Through a combination of strategic insights and applied exercises, the course explores fundamental concepts such as adaptive MT, terminology injection, prompt engineering for translation, RAG-based workflows, and model evaluation. Using OpenAI Agent Builder, participants will also implement a simulation of a hybrid MTQE-APE workflow.

By the end of the course, learners will be able to design, evaluate, and optimize AI-powered translation strategies aligned with real-world production needs.

Instructor: Andrés Romero Arcas
Andrés Romero Arcas is an AI and Language Technology Specialist with strong experience in NMT, LLMs, MTQE, and automated translation workflows. He focuses on practical, scalable AI solutions for localization teams and is known for explaining complex technology in a clear, actionable way.
Session 1: The AI Translation Landscape
Session 2: NMT Solutions: Fine-tuning
Session 3: Adaptive MT & Terminology Injection
Session 4: Machine Translation Quality Evaluation
Session 5: LLM Solutions for Translation
Session 6: Hybrid Workflow: MTQE & APE
Session 7: Hands-on Practice: Building MTQE-APE with OpenAI Agent Builder
Session 1

The AI Translation Landscape

Session 1. The AI Translation Landscape

1.1 Machine Translation fundamentals
1.2 Adoption challenges and strategic considerations
1.3 NMT vs. LLMs: capabilities, limitations, use cases
1.4 Market solutions for current translation bottlenecks
1.5 Choosing the right system for your workflow
Session 2

NMT Solutions:
Fine-tuning

Session 2. NMT Solutions: Fine-tuning

2.1 What is fine-tuning and why it matters
2.2 Overview of major NMT providers
2.3 Data selection, preparation, and cleaning
2.4 How NMT engines are trained, evaluated, deployed, and retrained
2.5 When to train and when not to train
SESSION 3

Adaptive MT & Terminology Injection

Session 3. Adaptive MT & Terminology Injection

3.1 How adaptive MT works and when to use it
3.2 Advantages and constraints
3.3 Terminology injection: strategies, mechanisms, and best practices
3.4 Recommendations for real-world implementation
Session 4

Machine Translation Quality Evaluation

Session 4. Machine Translation Quality Evaluation

4.1 Automatic quality evaluation (AQE): concepts and metrics
4.2 Reference-based vs. reference-free evaluation
4.3 Lexical and semantic metrics explained
4.4 Human evaluation methodologies
4.5 Deployment, monitoring, and continuous improvement
4.6 Recommendations for defining your NMT strategy
Session 5

LLM Solutions
for Translation

Session 5. LLM Solutions for Translation

5.1 How LLMs are developed and what differentiates them from NMT
5.2 Challenges for adoption (hallucinations, consistency, risk control)
5.3 Prompting for translation tasks
5.4 Retrieval Augmented Generation (RAG) for domain specificity
5.5 Fine-tuning LLMs for translation
5.6 Are LLMs the future of machine translation?
Session 6

Hybrid Workflow:
MTQE & APE

Session 6. Hybrid Workflow: MTQE & APE

6.1 Why hybrid workflows outperform traditional MT pipelines
6.2 MTQE + APE standard workflow
6.3 Quality estimation models and scoring
6.4 Automated post-editing mechanisms
6.5 Integrating human linguists into the workflow
6.6 Quality monitoring and management
6.7 Strategic best practices
Session 7

Hands-on Practice: Building MTQE-APE with OpenAI Agent Builder

Session 7. Hands-on Practice: Building MTQE-APE with OpenAI Agent Builder

7.1 Overview of Agent Builder architecture
7.2 Working with Agents, Variables, Set State, and If/Else
7.3 Designing logic for quality estimation and automatic post-editing
7.4 Building your own automated workflow simulation
Meet

Andrés Romero

Andrés is a proficient language technology expert with over a decade of experience in the Localization Industry.
Throughout his career, he has held diverse roles, such as CAT Tool Specialist, Localization Engineer and Operations Technology Coordinator, where he led a team of localization engineers.
Currently at Acolad, Andrés focuses on machine translation evaluation and engine training. He is also deeply involved in prompt engineering and Generative AI, proposing AI-driven driven solutions to deliver tailored, customer-centric solutions and to tackle challenges in Production.
Andrés is passionate about automating and optimizing processes to enhance productivity and efficiency, improving quality and integrating innovation into localization workflows.
Andrés Romero - Course Creator & Host
Created with