Senior Applied AI Researcher

About the Role You will bridge the gap between research and production systems that serve millions of users. In this role, you'll tackle fundamental ML questions while respecting scalability, latency, and real-world deployment constraints. You'll define your own research agenda aligned with product impact, taking projects from initial hypothesis through published results and production implementation. Working at the intersection of theory and practice, you'll push state-of-the-art boundaries in large-scale model training, efficient inference, or novel architectures—then see your innovations directly influence how our systems learn and adapt. Our Stack - Modern deep learning: PyTorch · JAX · Hugging Face · Weights & Biases - Cloud-native ML infrastructure on AWS/GCP with distributed training, experiment tracking, and production model serving - Full autonomy over your research stack—propose and adopt new tools that accelerate your work What You'll Do - Design and execute research projects that advance both scientific understanding and product capabilities, from formulating hypotheses through rigorous experimentation to production deployment - Develop novel machine learning methods using PyTorch or JAX, implementing prototypes that balance theoretical innovation with practical scalability constraints - Analyze complex technical tradeoffs across model architecture, training efficiency, and inference performance using statistical rigor and ablation studies - Publish findings in top-tier venues while translating research contributions into production systems that measurably improve user experience - Collaborate with ML engineers and product teams to identify high-impact research directions, providing technical guidance that shapes research roadmaps and product strategy - Mentor junior researchers and engineers on experimental design, statistical methodology, and the path from research idea to production-ready implementation - Drive reproducibility and scientific rigor across research initiatives through proper benchmarking, significance testing, and transparent documentation of experimental results What We're Looking For - PhD in Computer Science, Machine Learning, Statistics, or related field, OR 5+ years of industry research experience with strong publication record at top-tier venues (NeurIPS, ICML, ICLR, CVPR, ACL) - Deep expertise in PyTorch or JAX for implementing novel architectures from scratch—you've debugged gradient flows and custom CUDA kernels - Track record of taking research from hypothesis through rigorous experimentation to production deployment, balancing scientific rigor with practical constraints - Strong foundation in statistical analysis and experimental design—you know when results are significant, design proper ablation studies, and ensure reproducibility - Hands-on experience with distributed training across multi-GPU or multi-node setups, understanding tradeoffs between model and data parallelism - Ability to define your own research agenda aligned with business impact—comfortable proposing new directions and defending technical choices through clear written reasoning - Proven collaboration skills across disciplines—you've worked effectively with ML engineers on productionization and product teams to translate research into user value Nice to Have - Publications demonstrating innovation in your research area, whether academic papers, technical blog posts, or open-source contributions - Experience with the full ML deployment stack from experimentation (Weights & Biases, MLflow) to production inference systems - Familiarity with transformer architectures, large language models, or other cutting-edge AI paradigms relevant to applied problems Bonus Points - Open-source contributions to major ML frameworks (PyTorch, JAX, Hugging Face) or widely-used research codebases - Experience mentoring junior researchers or presenting at conferences - Track record of research that shipped to production and measurably improved product metrics