AI/ML Engineer - Time-Series

Role overview Join as one of the first machine learning hires to own the modeling stack for an AI-driven autoscaling product, from feature engineering on noisy real-world metrics to deployment and evaluation. Your work will directly determine how well the system learns workload patterns and recommends scaling configurations that customers run against in production.

Responsibilities - Build forecasting models for queue depth, task utilization, and processing rates across hourly, daily, and weekly cycles. - Develop anomaly detection that distinguishes predictable traffic peaks from genuine incidents. - Maintain the evaluation harness, including synthetic scenarios and historical replays that compare AI-tuned configurations against the customer's current setup. - Design the recommendation policy that translates a forecast into a concrete configuration diff covering thresholds, min/max bounds, and cooldowns. - Lay the foundation for the modeling platform that future ML engineers will depend on.

Requirements - Four or more years building production machine learning systems on time-series data, such as forecasting or anomaly detection. - Strong Python skills combined with a solid statistical foundation. - Comfortable defending modeling choices with numbers and evidence rather than intuition. - Track record of shipping models that run against real customer workloads.

Nice to have - Experience with observability data sources such as Prometheus or CloudWatch. - Familiarity with scheduler or autoscaling problems. - Deep-learning experience is welcome but not required.

Benefits and work setup - Fully remote, aligned with EU and Israel time zones. - Async-first culture that allocates time for research while keeping the bar on shipping production models.