AIML - Applied Research Engineer, Machine Translation

Apple; Aachen; Vor Ort; EUR 100.000 - 140.000; Vollzeit

Zusammenfassung

Apple’s Machine Translation team invites an Applied Research Engineer to advance state‑of‑the‑art language models. You will use SFT, RLHF and GRPO to improve translation quality across text and speech, building scalable pipelines from data collection to production rollout.

You’ll ship models touching hundreds of millions of users across Translate, Live Translation and system‑wide experiences, joining a collaborative, quality‑driven environment.

Qualifikationen

• Strong programming and software engineering skills (Python, C++, or equivalent), with hands‑on experience training and fine‑tuning large‑scale models.
• Experience building and optimizing machine translation, natural language processing, or related sequence‑to‑sequence systems using modern LLM architectures.
• Practical knowledge of LLM post‑training techniques, including SFT, RLHF, and GRPO or similar reward‑based optimization methods.
• Experience with large‑scale data processing frameworks (Spark, Dask, or equivalent) and synthetic data generation pipelines.
• Strong production mindset: ability to take models from research to reliable, customer‑facing deployment.
• Ability to manage complex processes across multiple stakeholders in a fast‑paced environment.
• Excellent communication skills and a proactive, collaborative approach to teamwork.
• Deep motivation to ship the best, most impactful products for Apple's customers.

Aufgaben

• Design and implement LLM fine‑tuning pipelines (SFT, RLHF, GRPO, and related RL‑based methods) tailored to machine translation quality objectives.
• Drive production model improvements end‑to‑end: from experimentation and offline evaluation through A/B testing and customer‑facing rollout.
• Generate and curate training data — both organic and LLM‑synthesized — to improve translation quality across text and audio input modalities and accelerate expansion into new languages.
• Develop and maintain large‑scale distributed training pipelines optimized for rapid iteration and reproducibility.
• Build robust tooling for automated quality checks, regression testing, and model benchmarking across existing and new language pairs.
• Define and track evaluation criteria and reward signals that reflect real‑world translation quality, enabling data‑driven decisions on model releases.
• Collaborate cross‑functionally to manage data assets, model versioning, and release schedules across a growing portfolio of languages and platforms.
• Stay current with the latest research in LLMs, MT, and RL‑based training methods, and rapidly prototype and integrate promising advances into production workflows.

Kenntnisse

Python

C++

LLM training

RLHF

SFT

GRPO

Ausbildung

Master’s degree or PhD in CS/EC Engineering or related

Tools

PyTorch

Cloud platforms (GCP/AWS)

Jobbeschreibung

Summary

Would you like to play a critical part in the next revolution of human‑computer interaction? The Apple Machine Translation team is building groundbreaking technology that enables connecting people across language barriers. We are looking for an Applied Research Engineer who is passionate about leveraging the latest advances in large language models and reinforcement learning to create, maintain, and ship scalable, high‑quality model assets across a multitude of languages — powering Apple's Machine Translation products such as the Translate App, Safari web translation, system‑wide translation, and Live Translation, powered by Apple Intelligence.

Description

Apple's Machine Translation is deeply embedded across the iOS, iPadOS, macOS, and watchOS ecosystems: from the Translate App that bridges communication across languages, to Live Translation, powered by Apple Intelligence, which enables seamless, real‑time translation experiences across calls, messages, and everyday interactions. As LLMs redefine what is possible in natural language understanding and generation, this role sits at the intersection of cutting‑edge research and real‑world product impact.

You will apply and advance modern training paradigms, including SFT, RL‑based fine‑tuning, and preference optimization, to push translation quality to new heights across text and speech modalities. You will own and improve end‑to‑end model development pipelines, from data acquisition and synthetic data generation through training, evaluation, and production rollout.

You will be part of a motivated and dynamic team responsible for shipping models that reach hundreds of millions of users, with a relentless focus on quality, efficiency, and continuous improvement.

Responsibilities

• Design and implement LLM fine‑tuning pipelines (SFT, RLHF, GRPO, and related RL‑based methods) tailored to machine translation quality objectives.
• Drive production model improvements end‑to‑end: from experimentation and offline evaluation through A/B testing and customer‑facing rollout.
• Generate and curate training data — both organic and LLM‑synthesized — to improve translation quality across text and audio input modalities and accelerate expansion into new languages.
• Develop and maintain large‑scale distributed training pipelines optimized for rapid iteration and reproducibility.
• Build robust tooling for automated quality checks, regression testing, and model benchmarking across existing and new language pairs.
• Define and track evaluation criteria and reward signals that reflect real‑world translation quality, enabling data‑driven decisions on model releases.
• Collaborate cross‑functionally to manage data assets, model versioning, and release schedules across a growing portfolio of languages and platforms.
• Stay current with the latest research in LLMs, MT, and RL‑based training methods, and rapidly prototype and integrate promising advances into production workflows.

Minimum Qualifications

• Strong programming and software engineering skills (Python, C++, or equivalent), with hands‑on experience training and fine‑tuning large‑scale models.
• Experience building and optimizing machine translation, natural language processing, or related sequence‑to‑sequence systems using modern LLM architectures.
• Practical knowledge of LLM post‑training techniques, including Supervised Fine‑Tuning (SFT), Reinforcement Learning from Human Feedback (RLHF), and Group Preference Optimization (GRPO) or similar reward‑based optimization methods.
• Experience with large‑scale data processing frameworks (Spark, Dask, or equivalent) and synthetic data generation pipelines.
• Strong production mindset: ability to take models from research to reliable, customer‑facing deployment.
• Ability to manage complex processes across multiple stakeholders in a fast‑paced environment.
• Excellent communication skills and a proactive, collaborative approach to teamwork.
• Deep motivation to ship the best, most impactful products for Apple's customers.

Preferred Qualifications

• Master’s degree or PhD in Computer Science, Electrical and Computer Engineering, or related field.
• Experience in applied machine learning or software engineering, with demonstrable impact on shipped products or systems.
• Hands‑on experience with deep learning frameworks (PyTorch or equivalent) and large‑scale model training.
• Familiarity with reward modeling, preference data collection, or RL‑based fine‑tuning for language models is a strong plus.
• Distributed and cloud computing experience (GCP, AWS, or equivalent) is a plus.
• Experience with speech translation or multimodal models is a plus.

At Apple, we’re not all the same. And that’s our greatest strength. We draw on the differences in who we are, what we’ve experienced, and how we think. Because to create products that serve everyone, we believe in including everyone. Therefore, we are committed to treating all applicants fairly and equally. We will work with applicants to make any reasonable accommodations.

At Apple, we believe accessibility is a fundamental human right. You’ll find that idea reflected in everything here — in our culture, our benefits and our digital tools. By welcoming as many perspectives as possible, we help you build a career where you feel like you belong.

Learn about accessibility in Apple’s workplace.

Role Number: 200670490-5487