AI Engineer

At Opsfleet, we don’t just “wrap” LLMs; we build autonomous, production-grade agentic systems. We are seeking an Engineer who understands that the future of AI isn’t a single prompt, but a complex, stateful ecosystem of collaborating agents. You will be responsible for designing the cognitive architectures that allow multiple agents to negotiate, self-correct, and execute complex workflows in production., * Agentic System Design: Architect and deploy complex Multi-Agent Systems (MAS) that move beyond basic RAG into autonomous reasoning, planning, and execution.

• Context Orchestration: Design and implement dynamic Context Management strategies, including sliding windows, semantic summarization, and token-efficient pruning to maintain long-term coherence without exceeding model limits.
• Tool-Use & API Synthesis: Expertise in Function Calling and Tool Definition; specifically creating standardized interfaces using MCP for agents to interact with external APIs, databases, and local environments.
• Orchestration & Frameworks: You will be responsible for designing and implementing robust state-handling for long-running agentic tasks using LangChain, LangGraph, Agno, Strands, ADK, etc.
• Safety & Guardrails: Design and implement Production Safety Measures, including output sanitization, hallucination triggers, and autonomous circuit breakers to ensure reliability in high-stakes environments.
• Memory & Personalization: Build sophisticated memory layers that manage User Preferences and historical context, ensuring agents provide personalized, context-aware interactions over time.
• System Evaluations: Define and implement rigorous requirements for Agentic Evaluation, benchmarking trajectory consistency, tool-calling accuracy, and reasoning logic., * Define Agentic Autonomy: You aren’t just following tutorials; you are architecting the blueprint for how autonomous systems reason, collaborate, and execute. You will set the industry standards for agentic maturity, moving beyond simple task-execution into complex, self-correcting cognitive architectures.
• Complex Projects: Work on everything from internal coding agents to massive client-facing multi-agent orchestrators.
• Remote-First: A culture built by engineers, for engineers

Employment: Full-time position (40 hours/week)

Benefits: * Potential career paths * Collaborative remote team environment * Opportunity to contribute to open-source GenAI initiatives and research publications

We are seeking a talented and experienced AI Engineer with a background in Data Science, Machine Learning, Backend development, or equivalent to join our professional services team of over 60 engineers, headquartered in Israel. The position is available full-time. This remote-first role focuses on Generative AI, specifically working with advanced Large Language Models (LLMs) and building AI agents. You’ll work with state-of-the-art foundation models through platforms like AWS Bedrock and GCP Vertex AI to develop innovative GenAI applications and scalable AI infrastructure for our clients., * Software Design: 5+ years of Python experience with a focus on Modular Architecture and clean code principles.

• Agentic Frameworks: Expert-level proficiency in modern orchestration frameworks (LangChain, LangGraph, Agno, Strands, ADK, etc.).
• Cloud Infrastructure: Deep experience with AWS Bedrock or GCP Vertex AI for enterprise-scale model deployment and agent hosting.
• Communication Standards: Hands-on experience with MCP for connecting agents to diverse data sources and tools, and A2A for cross-agent collaboration.
• Vector & Graph Databases: Proficiency in using Knowledge Graphs and vector storage to ground agentic reasoning and provide structured organizational memory., * Agent Engine Mastery: Direct experience working with AgentCore architectures and native Agent Engines (AWS Bedrock Agents / GCP Vertex AI Agent Engine).
• Multi-Modal Expertise: Experience building agents capable of synthesizing information across Text, Vision, and Video sources.
• The “Agentic Mindset”: You can explain the difference between a linear chain and a directed acyclic graph (DAG), and you know exactly when an agent is the right tool vs. a simple workflow.
• The Reviewer’s Eye: Experience conducting deep technical reviews of AI implementations, identifying architectural bottlenecks or “lazy” agent logic.
• Production Battle Scars: You’ve dealt with agents getting stuck in infinite loops or hallucinating tool-calls, and you know how to build the infrastructure to prevent it.