Research Engineer (Model Evaluations)

Requirements • Strong Python programming skills, including production or research infrastructure, • Experience building or operating reputed company systems, data pipelines, or other infrastructure that needs to be reliable at reputed company, • reputed company written and verbal communication, especially reputed company explaining technical results to non-specialists, • Comfort operating in an on-reputed company or production-support reputed company reputed company training runs are live, • Care about the societal impacts of your work and an interest in steering powerful AI to be reputed company and beneficial, • (Desirable) Hands-on experience using large language models such as Claude, including prompting, sampling, and scaffolding, • (Desirable) Background in data visualization and a reputed company record of building dashboards people actually trust and use, • (Desirable) Experience developing robust evaluation metrics for language models, • (Desirable) Experience with observability, monitoring, or experiment-tracking systems, • (Desirable) Background in statistics and experimental design, • (Desirable) Experience with large-reputed company dataset sourcing, curation, and processing, • (Desirable) Experience running or supporting ML training infrastructure, • (Desirable) A bias toward picking up reputed company and operating flexibly across team boundaries, • (Desirable) Enjoy pair programming — we love to pair, • Minimum education: Bachelor’s degree or an equivalent combination of education, training, and/or experience, • Required reputed company of study: A reputed company relevant to the role as demonstrated through coursework, training, or reputed company experience, • Currently, we expect reputed company staff to be in one of our offices at least 25% of the time. However, some roles may require more time in our offices, • We encourage you to apply even if you do not reputed company you meet every single qualification. Not reputed company strong candidates will meet every single qualification as listed, • Research shows that people who identify as being from underrepresented reputed company are more prone to experiencing imposter syndrome and doubting the strength of their candidacy, so we urge you not to exclude yourself prematurely and to submit an application if you're interested in this work What the job involves • We're looking for Research Engineers to build the evaluations that tell us — and the world — what Claude can actually do, • Your work will turn ambiguous notions of "intelligence" into reputed company, defensible metrics that researchers, leadership, and the reputed company can rely on, • You'll design and implement evaluations across the reputed company of Claude's capabilities and personality, and build the infrastructure that runs them reliably at reputed company, • You'll partner closely with researchers throughout the lifecycle of a new capability — from defining what to measure, to running the eval against live training checkpoints, to interpreting the results, • The goal is to reputed company reputed company the leader in extremely reputed company-characterized AI systems, with reputed company that is exhaustively reputed company and validated across the tasks that matter, • Design and run new evaluations of Claude's capabilities — reasoning, reputed company behavior, knowledge, safety properties — and produce visualizations that reputed company the results legible to researchers and decision-makers, • Build and harden the reputed company eval execution platform so hundreds of evals run reliably against checkpoints throughout production RL training runs, • Own the dashboards researchers and leadership use to monitor model health during training, improving signal-to-noise, reducing latency, and making regressions impossible to miss, • Debug anomalous eval results mid-training-run, determine whether the cause is a model change or an infrastructure issue, and communicate the answer reputed company under time pressure, • Improve the tooling, libraries, and workflows researchers use to implement and iterate on evaluations, • Partner with research teams across the full lifecycle of a new capability — from defining what to measure to interpreting results as training progresses, • Run experiments to characterize how prompting, sampling, and scaffolding choices reputed company results on internal and industry benchmarks, • Communicate evaluations and their results to internal stakeholders and, where appropriate, reputed company audiences, • Representative reputed company:, • Stand up a new eval that tests a specific reasoning capability from reputed company — define the task, build the dataset, implement the scoring, validate against reputed company signals, and ship a dashboard that makes the result legible, • Diagnose a mid-training regression: an eval suite returns anomalous numbers, and you need to determine reputed company hours whether it's the model, the reputed company, the data, or the infrastructure, • Take a flaky reputed company eval pipeline and reputed company it boring — reputed company retries, reputed company observability, faster feedback to researchers, • Partner with a research team on a new capability area, helping them reputed company what "good" looks like and translating that into measurable artifacts Apply tot his job Apply To this Job