About this role
The Compliance Engineering team at Goldman Sachs is seeking an experienced AI/ML Engineer at the Vice President level to build and deliver high-impact AI/ML solutions for complex regulatory and reputational risk management. This role involves working with petabyte-scale structured and unstructured data to develop vertical AI agents and advanced models.
Responsibilities
- Design and develop end-to-end AI/ML solutions for compliance, including GenAI-driven applications, agentic frameworks, and RAG pipelines.
- Implement model fine-tuning, prompt engineering, and algorithmic experimentation to solve novel business challenges.
- Develop, test, and maintain production-ready code while driving technical deliverables from design to implementation.
- Collaborate with compliance officers and legal counsel to translate business requirements into technical solutions.
- Participate in code reviews and promote best practices in MLOps, version control, and documentation.
Required Qualifications
- Bachelor's, Master's, or PhD in Computer Science, Machine Learning, Mathematics, or a related field.
- Minimum 6 years of industry experience for Bachelor's/Master's degree holders, or 4 years for PhD holders, with a focus on Language Models.
- Strong foundation in machine learning and deep learning architectures (Transformers, RNNs, CNNs).
- Proficiency in Python and frameworks such as TensorFlow, PyTorch, and Hugging Face.
- Expertise in GenAI techniques (RAG, fine-tuning, prompt engineering, AI agents) and MLOps (Docker, Kubernetes, CI/CD).
Preferred Qualifications
- Experience with agentic frameworks like LangChain or AutoGen.
- Knowledge of performance optimization for real-time inference (quantization, pruning, knowledge distillation).
- Familiarity with model interpretability, data governance, and financial compliance regulations.
Skills for this role
PythonTensorFlowPyTorchHugging Face Transformersscikit-learnGenerative AIRAGPrompt EngineeringAI AgentsMachine LearningDeep LearningMLOpsDockerKubernetesCI/CDLangChainAutoGenModel Fine-tuningVector DatabasesDistributed Systems