LLM Fine-Tuning / Model Engineer Financial Services

OverviewWe are looking for an experienced LLM Fine-Tuning & Model Engineer to adapt, evaluate, optimize, and deploy Large Language Models for enterprise AI use cases. The ideal candidate will have strong expertise in model fine-tuning, evaluation, optimization, and deployment of open-weight and custom foundation models using modern ML and GenAI technologies.Key ResponsibilitiesFine-tune and optimize Large Language Models using domain-specific datasets.Design and implement training pipelines utilizing LoRA, QLoRA, PEFT, and Supervised Fine-Tuning (SFT).Perform model evaluation, benchmarking, and performance analysis.Build and maintain RAG-based AI systems.Deploy and monitor models using Azure ML and cloud-native infrastructure.Optimize model performance, inference costs, latency, and scalability.Conduct experimentation around model distillation and alignment techniques.Collaborate with AI Engineers and Data Scientists to improve model quality.Ensure model governance, security, and responsible AI standards.Required SkillsAdvanced Python programming skills.Deep understanding of NLP, LLMs, and Transformer architectures.Hands-on experience with:LoRA / QLoRAPEFTSupervised Fine-Tuning (SFT)Model DistillationModel Evaluation FrameworksExperience with Hugging Face ecosystem.Experience deploying open-source models such as Llama, Mistral, Falcon, Qwen, or similar models.Experience with Azure Machine Learning and Azure OpenAI.Strong understanding of RAG architectures and vector databases.Preferred SkillsExperience with distributed training techniques.Knowledge of reinforcement learning from human feedback (RLHF).Exposure to Agentic AI frameworks.Financial services or banking domain experience.MLOps and CI/CD experience.QualificationsBachelor's or master's degree in computer science, Machine Learning, Artificial Intelligence, or related field.4+ years of Machine Learning or AI engineering experience.Proven experience fine-tuning and deploying LLMs in production environments.