Staff Machine Learning Engineer

Zscaler processes an enormous volume of security telemetry through what it describes as the world's largest security data lake, and machine learning is how it separates real attacks from ordinary traffic at that scale. This Staff role is more infrastructure heavy than most machine learning jobs: the posting describes building and maintaining large scale distributed systems that support the entire pipeline, from data collection and feature engineering through model training, evaluation, deployment and real time serving. The stated problems are threat detection and business intelligence, worked on directly with data scientists, product management and product engineering. You would also write functional specifications, assess requirements and scheduling, and stay involved through development, debugging and support. Staff at Zscaler is a senior individual contributor level, so the expectation is technical depth and ownership rather than managing people. The five year requirement sits above the company's other machine learning opening in this edition.

What the posting requires

- 5+ years as a software engineer or machine learning engineer
- Experience applying machine learning to improve key business metrics, optimising models for both accuracy and latency
- Experience researching newer techniques and innovating to build ML based solutions
- Experience with the algorithms behind data collection and processing infrastructure, model training, and serving platforms
- Deep expertise with AWS

Counts as a bonus, not a requirement

- A proven track record designing, building and shipping end to end applications at scale
- Experience identifying and solving problems in data using machine learning systems
- Familiarity with multi agent systems, agent architectures, prompt engineering and current AI frameworks

What you would actually be doing

- Building and maintaining large scale distributed systems covering the whole pipeline: data collection, feature engineering, model training, evaluation, deployment and real time serving
- Writing functional specifications, assessing task requirements and scheduling, and staying involved through development, debugging and support
- Solving problems such as threat detection and business intelligence in direct collaboration with data scientists, product management and product engineering

Location and working pattern

Based in Bangalore, reporting to the Manager of Machine Learning Engineering. Zscaler operates a hybrid working model.

A good fit if

You have several years of production machine learning and want an adversarial domain. Threat detection behaves very differently from the recommendation and forecasting problems most machine learning engineers see, because the data changes in response to your model.

Think twice if

You prefer modelling to infrastructure. A large part of this role is the distributed systems underneath the models, and the posting leads with that rather than with modelling technique.