AIML - Senior Machine Learning Research Engineer, LLM Post-training (Multilinguality)

Socket.dev; Cupertino (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Socket.dev in the United States is seeking a Machine Learning Engineer to advance language identification capabilities. You will design and train models that classify languages across scripts, regions, and user contexts, and you will contribute to robust, scalable multilingual features used by millions.

You will work across the full model lifecycle—from data curation and model design to evaluation, optimization, and deployment—collaborating with researchers, engineers, and product teams to

Qualifications

• 2+ years of ML/NLP experience with neural models.
• Proficient in Python and PyTorch (or TensorFlow/JAX).
• BS or MS in CS/ML/Computational Linguistics or equivalent.
• Experience with multilingual datasets.

Responsibilities

• Develop language identification models and benchmarks.
• Curate and preprocess multilingual data.
• Evaluate models and iterate on architecture and training.
• Collaborate with researchers and engineers to deploy scalable solutions.

Skills

Python

Deep learning

NLP

Model evaluation

Communication

Team collaboration

Education

Bachelor's or Master's in CS/ML/Computational Linguistics

Tools

PyTorch

TensorFlow

JAX

Hugging Face

Bash/Linux

Job description

The Multilingual Intelligence team is looking for a machine learning engineer to build the next generation of text language identification systems. You will develop models that accurately detect and classify languages across diverse scripts, regions, and user contexts — forming the backbone of multilingual features used by millions of people worldwide. You will work alongside a team of world-class experts to explore novel modeling approaches, data strategies, and evaluation methodologies that push the boundaries of what's possible in language detection at scale. Passionate about Natural Language Processing, multilingual systems, and building ML that works for everyone regardless of the language they speak? Join us to make every device fluent in every language.

Description

Text language identification is a foundational capability that powers multilingual experiences across products — from translation and search to content recommendation and accessibility. As the number of supported languages grows and user expectations rise, the challenge is building models that are not only accurate but also fair, robust, and efficient across the full spectrum of the world's languages. We are looking for a machine learning engineer passionate about building high-quality, inclusive language technology. As a member of the team, you will work across the full model lifecycle — from data curation and model design to evaluation and deployment. You will collaborate with researchers, engineers, and linguists to ensure our language identification systems perform reliably for users everywhere, regardless of how they write or what language they use. The successful candidate should be a strong team player with excellent oral and written communication skills and a genuine passion for building ML systems that work for the world's linguistic diversity.

Minimum Qualifications

• 2+ year experience in machine learning, NLP, or related fields, with hands-on experience training and evaluating neural models.
• Proficient programming skills in Python and at least one major deep learning framework such as PyTorch, TensorFlow, or JAX.
• Bachelor's or Master's degree, or equivalent practical experience, in Computer Science, Machine Learning, Computational Linguistics, or a related technical field.
• Experience working with multilingual, multi-script, or cross-lingual datasets.

Preferred Qualifications

• Experience with text classification, language identification, dialect identification, or similar NLP tasks involving multiple languages.
• Familiarity with embedding models, sentence representations, or contrastive learning methods.
• Understanding of model optimization techniques such as quantization, pruning, or knowledge distillation, and a general interest in efficient inference.
• Experience with low-resource languages, code-switching, or mixed-language input handling.
• Familiarity with Hugging Face ecosystem (transformers, tokenizers, datasets) and modern NLP pipelines.
• Ability to formulate a problem, design experiments, and implement end-to-end solutions in Python and Bash.
• Strong communication skills and a passion for working cross-functionally across Research, Engineering, and Product teams.