AI platform for automotive cybersecurity and vehicle data management
Upstream builds a data platform purpose-built for connected vehicles, using Python, Pandas, Polars, scikit-learn, Kafka, and Spark to centralize fragmented mobility data and power AI-driven security and quality applications. The tech stack reveals a mature data-science organization (MLflow, dbt, Airflow, Prefect for orchestration) layered with infrastructure-as-code tooling (Terraform, Terragrunt, Kustomize, Flux, ArgoCD), now adopting AWS, Azure, and GCP in parallel—suggesting multi-cloud deployment strategy. Hiring is balanced across data (4), sales (4), and engineering (2), with a senior-skewed leadership mix (6 directors + managers out of 16 open roles), pointing to scaling commercial motion alongside product maturity.
Upstream is an Israeli SaaS company (founded 2017, 51–200 employees) that delivers a cloud-based AI platform for automotive OEMs, suppliers, and mobility operators. The core product transforms distributed vehicle and IoT data into centralized, structured data lakes, then applies machine-learning models for cybersecurity detection and response (XDR), quality assurance, warranty analysis, and usage-based insurance. The company protects millions of vehicles across its customer base. Active projects span ML model productization, quality and cybersecurity detection, GenAI-based cloud data management, and platform automation—while scaling to handle integration with multiple vendor systems and growing the qualified pipeline in new geographies.
Python, Pandas, Polars, NumPy, scikit-learn, Kafka, Spark, Trino, dbt, Apache Airflow, Prefect, MLflow, Kubernetes, Terraform, and GitHub Actions. Currently adopting AWS, Azure, and GCP.
ML-driven detection and anomaly detection for automotive cybersecurity and quality; GenAI-based cloud data management; ML model productization and lifecycle automation; and expanding the platform to new use cases like warranty and recall analysis.
Other companies in the same industry, closest in size