Senior AI Engineer – LLM and Agent Systems

Description

Summary: Source Meridian seeks a Senior AI Engineer specializing in LLMs and agent systems to design, develop, and deploy intelligent applications for healthcare. Highlights: 1. Design and implement agentic workflows using LangGraph and LangChain. 2. Integrate and optimize LLMs, focusing on RAG and memory-managed architectures. 3. Collaborate on cutting-edge AI in a high-impact tech-health company. **We’re looking for a** **Senior AI Engineer – LLM and Agent Systems to join Source Meridian** ---------------------------------------------------------------------------------------------- **About Source Meridian** ------------------------- Source Meridian is a development software company that works to solve the industry’s most challenging problems in healthcare practices. We are laser focused on specific technologies in the healthcare and life science industries: Healthcare technology, artificial intelligence, and healthcare interoperability. **About the Role** ------------------ AI Engineer with hands\-on experience in building production\-grade artificial intelligence systems using large language models (LLMs), agentic frameworks, and modern data infrastructure. The ideal candidate designs, develops, and deploys intelligent applications that leverage LLM orchestration, retrieval\-augmented generation (RAG), and memory\-managed architectures. **What You’ll Do** ------------------ * Design and implement agentic workflows using LangGraph and LangChain, including multi\-step reasoning, tool usage, and human\-in\-the\-loop patterns. * Integrate LLMs (OpenAI, Anthropic, Google Vertex AI, open\-source models) via REST APIs and SDKs into scalable backend services. * Build and maintain RESTful APIs (FastAPI) to serve AI\-powered functionality to frontends and external consumers. * Architect and manage vector database solutions (Milvus) for semantic search and retrieval\-augmented generation (RAG). * Design context engineering strategies, including prompt templates, dynamic context window management, token optimization, and context compression techniques. * Implement short\-term memory (conversational buffers, sliding windows, summary memory) and long\-term memory (persistent vector stores, knowledge graphs, user profile stores) using MongoDB and vector databases. * Store and manage structured and unstructured data in MongoDB, designing schemas that support conversation history, user state, and agent checkpoints. * Evaluate and improve the quality of LLM responses through prompt engineering, few\-shot examples, guardrails, and automated evaluation pipelines. * Collaborate with DevOps teams to containerize and deploy AI services using Docker, Kubernetes, and CI/CD pipelines on AWS or GCP. **Required Qualifications** --------------------------- * Strong command of Python (3\+ years) * Demonstrable experience with LangChain and LangGraph (graph\-based agent orchestration, state management, conditional edges, parallel execution). * Solid understanding of the fundamentals of LLMs: tokenization, embeddings, temperature/sampling, RAG. * Hands\-on experience with vector databases and embedding models for semantic search and retrieval pipelines. * Experience designing and consuming REST APIs; knowledge of authentication. * Proficiency in MongoDB (document modeling, aggregation pipelines, indexing strategies). * Understanding memory architectures for conversational AI: summary memory, entity memory, and long\-term persistent stores. * Familiarity with context engineering: token budget management, hybrid search (sparse \+ dense). * English level: B2 or higher **Nice to Have** ---------------- * Experience with assessment frameworks (RAGAS, Langfuse (LangSmith), customized assessments). * Experience with streaming responses. **What We Offer** ----------------- Permanent contract Learning and continuous growth environment Benefits package focused on health and well\-being Competitive salary based on experience **Apply only if you reside in Colombia or Ecuador** At **Source Meridian**, you’ll be part of a high\-impact **tech\-health** company, building products that truly make a difference. If you meet the profile — or know someone who might be interested — **apply now!** We’d love to meet you