Machine Learning Engineer, Arthur AI

• 2+ years of professional experience in software engineering, ideally providing strong ML support (or equivalent)
• Deep experience writing production-grade Python
• Exposure to (or experience) ML frameworks, such as JAX, Tensorflow, or PyTorch
• Understanding of software engineering best practices, version control, and containerization
• Customer empathy and a willingness to engage directly with customers and stakeholders
• Experience deploying ML models to production
• Deep interest in LLMs (to start) and multimodal generative models (soon!), specifically using these to build software

Desirable

• Arthur is looking for a product-focused Machine Learning Engineer to join our team, to operate with autonomy and with the support of the team to lead and build out our product vision and deliver value to our customers
• Focus on ML, especially LLMs, model performance, security, and robustness: make LLMs and model-based systems respond accurately, quickly, and comprehensively to general and task-specific interactions by both natural and adversarial users – for example, checking veracity of responses, hardening against prompt injection, query-based routing to different models, and generation of evaluation datasets
• Build out new techniques and tooling in ML model monitoring: data drift detection, high-dimensional density estimation, time series anomaly detection; detection and mitigation of issues in fairness and bias in ML algorithms
• Run machine learning experiments independently and in collaboration with product teams, and communicate results to internal stakeholders as well as the broader ML community via publications in top-tier conferences, industry conference presentations, deep technical blogs, and events
• Drive value creation using our product with our customers in collaboration with our customer teams by engaging in targeted and focused short-term project work to understand, experiment, refine and deploy ML-enabled features for customer use-cases
• Stay up to date on the latest developments in the field – we operate in a nascent space with plenty of signal and plenty of noise coming from arXiv and open source project