Senior AI/ML Engineer – Conversational AI & Agentic Systems Client of LinkX LLC

Design and develop a dual AI assistant ecosystem for system automation and intelligence, leveraging on-premise LLMs, RAG pipelines, and multi-agent architectures. Required Technical Skills Proven hands-on experience in the following areas: 1. On-Premise LLM Deployment • Deploy and manage local LLMs • Use inference engines (vLLM, TensorRT-LLM) • Apply quantization techniques (GPTQ, AWQ, GGUF) • Fine-tune models (LoRA/QLoRA) 2. Retrieval-Augmented Generation (RAG) • Implement end-to-end on-prem RAG workflows • Manage and optimize vector databases • Work with local embedding models (e.g., sentence-transformers, BGE) • Optimize chunking, semantic search, and context window utilization 3. Multi-Agent Orchestration • Build multi-agent systems with coordinated workflows • Use LangChain, LangGraph, or custom orchestration frameworks • Manage agent state, error handling, and task delegation 4. Natural Language Processing • Multilingual NLP (English/Arabic) • Intent/entity recognition and text classification • Experience with Arabic models and tokenization 5. Backend Development • Strong Python development • Docker/Kubernetes deployment experience 6. On-Prem MLOps • Model deployment • Monitoring (latency, throughput, GPU usage) • Model versioning and A/B testing • Logging, tracing, and observability 7. Deploy & Optimize On-Prem LLMs • Support in Set up and maintain LLM infrastructure • Apply quantization and inference optimizations • Fine-tune models for internal configuration tasks 8. Build Multi-Agent System • Develop functional agents for automation and intelligence • Build orchestrator agents for coordination and workflow control • Implement analysis, validation, and monitoring agents 9. Platform Integration • Integrate agents with existing microservices • Implement tool/function calling • Build APIs for UI, automation, and external integrations 10. Monitoring & Optimization • Define and track performance metrics • Optimize GPU utilization and inference cost • Validate agent behavior, quality, and reliability