Job description

ABOUT ARENA INTELLIGENCE Arena is the platform for evaluating how AI models perform in the real world. Founded by researchers from UC Berkeley's SkyLab, we're on a mission to measure and advance the frontier of AI for real-world use, and to build the foundation for everyone to understand, shape, and benefit from it. Tens of millions of people use Arena each month to evaluate how frontier systems handle the work they actually do. The preferences they share power the most transparent, rigorous, and human-centered evaluations in AI. Leading AI labs, enterprises, and independent researchers rely on our work and open datasets to understand how models behave in real workflows: agentic coding, creative generation, professional productivity, and beyond. We go beyond leaderboards and decompose what human experience reveals about AI, so models advance toward the work people actually do. We're a team of researchers, academics, builders, and creatives from UC Berkeley, Google, Stanford, and DeepMind. We seek truth, move fast, and value craftsmanship, curiosity, and impact over hierarchy. We're building a company where thoughtful, curious people from all backgrounds can do their best work together, in an office culture that radiates excellence, energy, and focus. ABOUT THE ROLE Arena Intelligence is seeking a variety of machine learning scientists to help advance how we evaluate and understand AI models. You’ll help design and analyse experiments that uncover what makes models useful, trustworthy and capable through human preference signals. Your work will contribute to the scientific foundations of understanding AI at scale. This role is deeply interdisciplinary. You’ll work closely with engineers, product teams, marketing and the broader research community to develop new methods for comparing models, analyzing preference data, and disentangling performance factors like style, reasoning, and robustness. Your work will inform both the public leaderboard and the tools we provide to model developers. If you’re excited by open-ended questions, rigorous evaluation, and research that’s grounded in real-world impact, you’ll find a meaningful home here. We’re looking for: - Hands-on experience training large-scale models, including reward models, preference models, and fine-tuning LLMs with methods like RLHF, DPO, and contrastive learning. - Strong foundation in ML and statistics, with a track record of designing novel training objectives, evaluation schemes, or statistical frameworks to improve model reliability and alignment. - Fluent in the full experimental stack, from dataset design and large-batch training to rigorous evaluation and ablation, with an eye for what scales to production. - Deeply collaborative mindset, working closely with engineers to productionize research insights and iterating with product teams to align modeling goals with user needs. YOU’LL - Design and conduct experiments to evaluate AI model behavior across reasoning, style, robustness, and user preference dimensions - Develop new metrics, methodologies, and evaluation protocols that go beyond traditional benchmarks - Analyze large-scale human voting and interaction data to uncover insights into model performance and user preferences - Collaborate with engineers to implement and scale research findings into production systems - Prototype and test research ideas rapidly, balancing rigor with iteration speed - Author internal reports and external publications that contribute to the broader ML research community - Partner with model providers to shape evaluation questions and support responsible model testing - Contribute to the scientific integrity and transparency of the Arena Intelligence leaderboard and tools YOU’LL HAVE - PhD or equivalent research experience in Machine Learning, Natural Language Processing, Statistics, or a related field - Strong understanding of LLMs and modern deep learning architectures (e.g., Transformers, diffusion models, reinforcement learning with human feedback) Proficiency in Python and ML research libraries such as PyTorch, JAX, or TensorFlow - Demonstrated ability to design and analyze experiments with statistical rigor - Experience publishing research or working on open-source projects in ML, NLP, or AI evaluation - Comfortable working with real-world usage data and designing metrics beyond standard benchmarks - Ability to translate research questions into practical systems and collaborate across engineering and product teams - Passion for open science, reproducibility, and community-driven research. WHAT WE OFFER - We offer competitive compensation aligned to the markets where our team members are based. The compensation range for each role is listed on its posting and will depend on the candidate's experience and work location. - Comprehensive health and wellness benefits, including medical, dental, vision, and additional support programs. - The opportunity to work on cutting-edge AI with a small, mission-driven team - A culture that values transparency, trust, and community impact Come help build the space where anyone can explore and help shape the future of AI. Arena Intelligence provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, national origin, age, disability, genetics, sexual orientation, gender identity, or gender expression. We are committed to a diverse and inclusive workforce and welcome people from all backgrounds, experiences, perspectives, and abilities.

Requirements

Must-haves

If you’re excited by open-ended questions, rigorous evaluation, and research that’s grounded in real-world impact, you’ll find a meaningful home here. We’re looking for:
Hands-on experience training large-scale models, including reward models, preference models, and fine-tuning LLMs with methods like RLHF, DPO, and contrastive learning.
Strong foundation in ML and statistics, with a track record of designing novel training objectives, evaluation schemes, or statistical frameworks to improve model reliability and alignment.
Fluent in the full experimental stack, from dataset design and large-batch training to rigorous evaluation and ablation, with an eye for what scales to production.
Deeply collaborative mindset, working closely with engineers to productionize research insights and iterating with product teams to align modeling goals with user needs.
Design and conduct experiments to evaluate AI model behavior across reasoning, style, robustness, and user preference dimensions
Develop new metrics, methodologies, and evaluation protocols that go beyond traditional benchmarks
Analyze large-scale human voting and interaction data to uncover insights into model performance and user preferences
Collaborate with engineers to implement and scale research findings into production systems
Prototype and test research ideas rapidly, balancing rigor with iteration speed
Author internal reports and external publications that contribute to the broader ML research community
Partner with model providers to shape evaluation questions and support responsible model testing
Contribute to the scientific integrity and transparency of the Arena Intelligence leaderboard and tools
PhD or equivalent research experience in Machine Learning, Natural Language Processing, Statistics, or a related field
Strong understanding of LLMs and modern deep learning architectures (e.g., Transformers, diffusion models, reinforcement learning with human feedback)
Proficiency in Python and ML research libraries such as PyTorch, JAX, or TensorFlow
Demonstrated ability to design and analyze experiments with statistical rigor
Experience publishing research or working on open-source projects in ML, NLP, or AI evaluation
Comfortable working with real-world usage data and designing metrics beyond standard benchmarks
Ability to translate research questions into practical systems and collaborate across engineering and product teams
Passion for open science, reproducibility, and community-driven research.

Nice-to-haves

None flagged.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aigomachine learningmlpython

Platform: ashby
Effort: quick

Benefits

Benefits detected in the listing.

Dental

Vision

Wellness