[Hiring] Junior AI Research Engineer @Phrase

Role Description

Phrase's AI Research team builds and evaluates the proprietary translation models and agentic workflows that power the platform's language quality. This role sits inside that team as a hands-on contributor, running experiments, fine-tuning models, and building evaluation pipelines alongside senior researchers. You'll work across fine-tuning LLMs for domain-specific machine translation, multi-agent evaluation pipelines, automated quality profiling from style guides, and active learning through feedback loops on real customer data.

It's built for someone earlier in their research career with strong fundamentals who wants to grow their depth in applied NLP and machine translation inside a team that ships research into production.

What you'll be responsible for:

• Fine-tune and evaluate LLMs for translation and language-quality tasks, applying techniques such as LoRA, with support from senior researchers.
• Build evaluation pipelines for translation quality, applying automatic metrics (BLEU, TER, ChrF, MQM, COMET) and LLM-as-judge approaches, and contribute to their design.
• Build and evaluate agentic workflows spanning multi-step reasoning, tool use, and structured output generation.
• Implement and evaluate components within the team's model-development programme, working to a defined scope.
• Run reproducible experiments and document results and decisions clearly.
• Work with Product, Engineering, and Solutions to turn research into shippable capabilities.
• Benchmark frontier and open-source models against internal baselines.
• Support the evolution of quality-evaluation systems, including style-guide integration.
• Support data-pipeline and feedback-loop work for continuous model improvement.
• Take part in reading groups, deep dives, and code and research reviews.
• Share results early and build on the team's existing work.
• Contribute to hiring tasks as you grow into the role.

Qualifications

• Solid grounding in machine learning and NLP, with hands-on experience fine-tuning or working with LLMs or transformer-based models.
• Working knowledge of at least one relevant area: machine translation, MT evaluation, information extraction, text generation, or search.
• Exposure to, or strong interest in, agentic or LLM-orchestrated systems (ability to build them from first principles is desirable, not required).
• Experience running experiments and applying evaluation metrics, with awareness of both automatic and human-in-the-loop evaluation.
• Solid Python skills and experience with ML/NLP libraries such as Hugging Face or PyTorch.
• MSc in Computer Science, Computational Linguistics, or a related field, or a BSc with relevant experience, projects, or a strong portfolio.
• A PhD is welcome but not required (desirable).
• Internships, thesis work, or projects in NLP/MT, ideally with production or applied exposure (desirable).
• Familiarity with localization or MT concepts such as MQM, COMET, post-editing, or TMS integrations (desirable).
• Exposure to experiment tracking tools such as Weights & Biases, or agentic-pipeline frameworks such as PydanticAI or LangGraph (desirable).
• Publications or open-source contributions in NLP-related topics (desirable).
• Curious and coachable, with a habit of asking good questions and acting on feedback.
• Comfortable sharing results early, inviting challenge, and building on others' work.

Requirements

• Current tech stack includes Python, Hugging Face, PyTorch, Weights & Biases, PydanticAI, and LangGraph.

Benefits

• Work experience in a successful and growing global SaaS company.
• Be part of an international team in Europe, APAC, and the Americas.
• Expert colleagues in their field who are determined to build the best localization platform on the market.
• An agile work environment, where it is encouraged to take smart risks.
• A culture full of trust, support, and loyalty, where respectful and open feedback is valued.
• A positive, open-minded, and innovative atmosphere.
• Support in your professional development and personal career goals.

Company Description

Phrase is committed to ensuring equal pay for equal work and work of equal value between women and men. We welcome applications from all qualified candidates regardless of their sex, race or ethnicity, disability, religion/belief, sexual orientation, gender identity or age.