Tel Aviv-Yafo, Israel · hace 4 meses
Diseñar y construir flujos de trabajo algorítmicos de IA que convierten documentos y planos inmobiliarios no estructurados en contexto estructurado para agentes basados en LLM. También crear suites de evaluación e iterar sobre la calidad y el rendimiento en producción.
REAL is building an AI Execution Platform for real estate organizations. Today, the data required to run real estate is scattered across PDFs, spreadsheets, emails, drawings, public records, and disconnected systems, leading to preventable leakage, missed obligations and lost opportunities to improve performance. Used by leading enterprises, REAL converts this fragmented data into connected intelligence and automated action. With advanced AI, universal ingestion, and modular execution agents, REAL increases operational accuracy, uncovers financial discrepancies, and surfaces opportunities to optimize performance and enhance business outcomes. REAL Values Ownership: We take responsibility and move decisively. Clarity: We simplify complexity to deliver meaningful impact. Accuracy: Precision matters, in our product and in how we operate. Velocity: We work with urgency and intent. Partnership: We collaborate closely with our customers and with each other. Role Overview As an AI Algorithm Engineer at REAL, you’ll design and implement advanced algorithmic workflows that power our data digestion pipelines and conversational AI agent experience. You will: • Architect and build workflows that turn raw, unstructured data (e.g., PDF documents and ultra high-resolution architectural drawings) into meaningful, structured context for downstream analysis. • Combine techniques from classical NLP, computer vision, unsupervised learning, and graph theory to build robust end-to-end pipelines - not just standard LLM API calls, but intelligent decomposition, structuring, and context engineering. • Develop AI agent workflows that let customers explore, query, and reason about their data naturally and reliably. • Build strong evaluation infrastructure to benchmark and continuously improve both classical algorithmic components and LLM-based workflows. • Work primarily in Python, and collaborate across systems and services in TypeScript when needed. • Move fast from prototype to production, while maintaining correctness, scalability, and measurable quality. Requirements What You’ll Do Core Responsibilities • Design and implement end-to-end data digestion pipelines for complex unstructured and semi-structured inputs. • Integrate classical algorithms with LLM-centric workflows to produce high-quality contextual inputs for reasoning tasks. • Build systems for segmentation, structure extraction, semantic decomposition, embedding-based representations, and graph construction. • Develop agentic workflows that combine tools, retrieval, and reasoning for interactive customer experiences. • Design evaluation suites, datasets, metrics, and regression tests to measure quality, robustness, and performance over time. • Continuously refine workflows based on customer usage, failure analysis, and measurable improvements. What We're Looking For Required Qualifications • MSc (or equivalent experience) in Computer Science, Mathematics, Electrical Engineering, or a related quantitative field. • 5+ years of experience in data science / algorithm development / applied research engineering roles. • Strong experience building complex algorithmic workflows beyond model calls - including multi-stage pipelines, feature extraction, structure inference, and optimization. • Deep understanding of LLM digestion concepts: RAG, context engineering, chunking strategies, retrieval and ranking, tool usage patterns, and reliability techniques. • Experience designing and maintaining evaluation frameworks for both classical algorithms and LLM workflows (offline + online, regression, benchmarking). • Strong programming skills in Python; familiarity with TypeScript is a plus. • Ability to thrive in a high-ownership startup environment: ambiguity, speed, responsibility, and continuous learning. Nice To Have • Familiarity with vector databases and scalable retrieval architectures. • Familiarity with agent orchestration frameworks (e.g., LangChain) and tool-driven reasoning systems. • Experience deploying AI workflows into production with monitoring, guardrails, and iteration loops. Benefits Why Join REAL • You’ll work on problems that don’t have a pre-scripted solution - it’s a privilege to be among the first to tackle them. • High ownership and direct impact on the core product and architecture. • A fast iteration environment where strong engineers ship meaningful work quickly. • A chance to grow alongside the AI revolution - adapting, learning, and building what’s next.
Publicada el 28 jun 2026
