Parallel Wireless · Tel-Aviv
Data & AI399-DevelopmentHybridFull Timeposted 6d ago
Design, build, and improve machine learning models and graph/statistical algorithms — including clustering, anomaly detection, and time-series modeling and forecasting — to drive automated network optimization..
Build a real, repeatable experimentation and model-deployment workflow (e.g., using MLflow or comparable tooling), taking models from notebook to production.
Work directly with the near-real-time engineering team to identify where today's rule-based, threshold-driven decisions can be replaced by learned models that adapt to real network conditions.
Mine large-scale time-series and topology data (stored in Elasticsearch, InfluxDB, and MongoDB) to uncover patterns in network behavior at the scale of thousands of cells and large user populations.
Define and track quantitative success metrics so every model shipped can be proven to actually improve network outcomes, not just deployed and forgotten.
Present findings and roadmap recommendations to engineering and business leadership.
5+ years of applied data science / machine learning experience, including graph algorithms, clustering, anomaly detection, or time-series modeling.
Strong Python skills; comfort working alongside Go-based production services.
Experience with large-scale time-series and document data stores (Elasticsearch, InfluxDB, MongoDB, or comparable).
Experience with ML experiment tracking and deployment tooling (MLflow or equivalent) and with streaming data pipelines (Kafka or comparable).
Excellent communication skills — the ability to turn open-ended "why is the network behaving this way" questions into a shipped, measurable model.
M.Sc. in Electrical Engineering, Computer Science, or Software Engineering.
Background in telecom, RF, or wireless networking (handovers, KPIs such as RSRP or PRB utilization, cell topology).
Experience turning a hand-tuned, rule-based system into a learned model running in a live production environment.
Experience with distributed or streaming compute frameworks.