Staff+ Software Engineer, Account Abuse (Machine Learning)
About this role
Bad actors constantly attempt to drain computing capacity and bypass safeguards, creating an ongoing need to detect and stop account abuse at scale. Anthropic's Account Abuse team tackles this challenge by developing machine learning systems that ensure resources are allocated fairly while minimizing the footprint of malicious users. Engineers in this position focus on classical machine learning using structured and behavioral data, building feature computation platforms and automated tooling that incorporates Claude to accelerate model development.
Candidates who succeed here possess a strong background in full stack machine learning and understand how to anticipate the strategies of opponents. Because false positives directly lock out legitimate customers, precision, measurement, and safe rollout procedures are critical aspects of the daily work. The environment suits professionals who care deeply about building robust production systems and protecting the integrity of large-scale AI infrastructure.
What you will do
- Build and operate a feature computation platform that serves both model training and real-time scoring with low-latency online retrieval.
- Train, evaluate, and deploy models that detect account-level abuse and fraud in both offline and online environments.
- Develop automation tooling for the model development lifecycle, including utilizing Claude to speed up feature development and evaluation.
- Establish backtesting, shadow deployment, and staged rollout as the default path to production with monitoring for drift and adversarial adaptation.
- Collaborate with data scientists and the Policy and Enforcement team to improve label coverage and quality.
- Partner with product and platform teams to integrate model decisions without negatively impacting system latency, stability, or overall architecture.
What you need
- Demonstrate proficiency in Python and SQL.
- Bring experience training machine learning models and deploying them to production environments.
- Build data pipelines utilizing batch processing engines like Spark or Beam alongside workflow schedulers like Airflow.
- Possess a working understanding of point-in-time correctness and training or serving skew, along with methods to prevent both.
- Communicate effectively with the ability to explain technical tradeoffs to non-technical stakeholders.
Nice to have
- Experience building or operating a feature platform such as Chronon, Feast, or Tecton
- Experience with stream processing engines such as Flink, Beam / Dataflow, or Kafka Streams
- Experience training ML models in a production setting with demanding serving requirements such as fraud, risk, or ranking
- Experience with tree-based models on tabular data
- Experience building unsupervised, clustering-based or graph-based detection systems
Benefits
- Competitive compensation and benefits
- Optional equity donation matching
- Generous vacation and parental leave
- Flexible working hours
Skills and tech stack
Questions candidates ask
What is the remote work and office policy?
Anthropic operates on a location-based hybrid policy. Staff are currently expected to be in one of the offices in San Francisco or New York City at least 25% of the time, though some roles may require more frequent in-office presence.
What is the salary range for this role?
The annual base salary for this position ranges from $320,000 to $485,000 USD.
What kind of machine learning background is required?
This role focuses on classical machine learning using structured and behavioral data rather than deep learning or LLM internals. Candidates need experience training and shipping models where stakes are real, alongside strong production engineering skills.
Does Anthropic provide visa sponsorship?
Yes, the company sponsors visas when possible. If an offer is made, the team makes every reasonable effort to secure a visa with the assistance of a retained immigration lawyer.