AI Agent Engineer — LLMs, Reinforcement Learning, RAG & Autonomous Systems

### Project Overview We are seeking an experienced AI Engineer, ideally with a Master’s or PhD in Computer Science, Artificial Intelligence, Machine Learning, Data Science, or a related field, to design and develop a reliable autonomous AI agent. This is not a basic chatbot or a simple LLM API integration project. We need a hands-on engineer who understands machine learning fundamentals, can build sophisticated agentic workflows, and knows how to use feedback mechanisms and applied reinforcement learning concepts to improve an agent’s performance over time. The ideal candidate combines a strong academic or theoretical foundation with the practical engineering skills required to build scalable, production-ready AI systems. ### The Challenge Large Language Models are highly capable, but they often struggle to complete complex, multi-step tasks consistently. They may lose context, repeat actions, hallucinate information, or drift away from the original objective. Your role will be to develop an agent that can: * Interpret and break down high-level goals * Create and execute multi-step plans * Interact with external tools and data sources * Evaluate the quality of its own outputs * Learn from environmental feedback * Revise its strategy when an approach is unsuccessful ### Key Responsibilities **Agent Design and Architecture** Design an autonomous agent capable of planning, reasoning, executing tasks, maintaining state, and using external tools such as APIs, databases, web search, and internal services. **LLM Workflow Orchestration** Develop reliable workflows using leading proprietary or open-source language models. Experience with frameworks such as LangGraph, AutoGen, CrewAI, or similar orchestration tools is highly valuable. **Feedback Loops and Applied Reinforcement Learning** Implement self-evaluation, reflection, reward scoring, retry logic, and other feedback mechanisms that allow the agent to recognize errors and improve its decision-making. **Retrieval-Augmented Generation** Build accurate and scalable RAG pipelines that provide the agent with relevant, current, and verifiable information from our documents, databases, and internal datasets. **System and Backend Integration** Write clean, modular, well-documented Python code and integrate the agent with our existing backend systems, APIs, and infrastructure. **Testing and Evaluation** Create evaluation frameworks and performance metrics to measure task completion, factual accuracy, reliability, latency, and cost. ### Required Experience * Master’s or PhD in Computer Science, AI, Machine Learning, Data Science, or equivalent advanced hands-on experience * Strong understanding of machine learning, neural networks, and modern AI systems * Advanced Python development skills * Experience with PyTorch, TensorFlow, Scikit-learn, or similar ML libraries * Proven experience working with LLMs, prompt design, structured outputs, tool calling, and context management * Experience developing AI agents or multi-step LLM workflows * Understanding of reinforcement learning concepts, reward functions, and feedback-based optimization * Experience building RAG systems using vector databases, embeddings, reranking, and document retrieval * Ability to write clean, maintainable, documented, and production-ready software ### Ideal Candidate You may be a strong fit for this project if: * You understand the theory and mathematics behind modern AI but prefer building practical systems * You have experience troubleshooting agents that hallucinate, repeat actions, or become stuck in execution loops * You know how to balance model quality, speed, reliability, and operating cost * You can translate complex AI concepts into clear and practical explanations * You communicate consistently and document your technical decisions * You are comfortable experimenting, measuring results, and improving systems through iteration ### When Applying Please include: * A brief overview of your experience building LLM-based agents * Examples of relevant AI, RAG, or autonomous-agent projects * The models, frameworks, and vector databases you have worked with * Your experience implementing evaluation or feedback mechanisms * Your availability and estimated approach for the initial development phase We are looking for someone who can take ownership of the architecture and help us build an AI agent that is reliable, measurable, and capable of improving over time. This version also adds testing, evaluation metrics, and application instructions to help attract stronger candidates.