Insurance and investment platform modernizing legacy operations with AI and data infrastructure
Harel is a publicly traded Israeli insurance and investment group managing over NIS 250 billion in assets, now building AI and automation capabilities at scale. The tech stack reveals a finance-and-data-led modernization: Python, LangChain, LangGraph, PyTorch, and TensorFlow anchor an emerging AI layer; Azure, Databricks, and Apache Spark power data pipelines; and RPA tools (UiPath, Power Automate) target process automation. Current hiring is heavily weighted toward finance and data roles (18 of 26 open positions), with active projects spanning LLM agents, anomaly detection in documents, and end-to-end business process improvement — suggesting a focused push to automate legacy manual workflows and embed AI into underwriting and portfolio management.
Harel Insurance, founded in 1935 and headquartered in Ramat Gan, Israel, is one of the country's largest insurance and financial services groups with 1,001–5,000 employees. The company offers comprehensive insurance products (health, life, business, motor), pension and mutual funds, and investment portfolio management, operating as a global partner to major insurers including Zurich, Allianz, AXA, and Chubb. In 2015, management launched a digital transformation strategy focused on process digitization, data-driven service delivery, and analytics—a shift now reflected in their active development of AI evaluation frameworks, document processing systems, and cross-organizational automation initiatives. Core operational pain points center on document handling, anomaly detection, SOX compliance, and complex manual processes like daily bank adjustments and investment return calculations.
Harel's stack includes LangChain, LangGraph, PyTorch, TensorFlow, scikit-learn, MLflow, and OpenAI. Active projects include LLM agents, RAG systems, and smart ML/DL models for decision-making and anomaly detection.
Harel uses Azure, AWS, GCP, and Databricks. The stack emphasizes Azure and Databricks for workflows and data pipelines, alongside Apache Spark for distributed processing.
Other companies in the same industry, closest in size