Fulltime :: Microsoft Foundry Specialist — AI/ML Engineer

Ampstek is a small, Agile/Scrum team delivering AI and digital transformation solutions to enterprise clients. They are seeking a Microsoft Foundry Specialist with expertise in machine learning and experience on the Microsoft Foundry platform to design, build, and deploy AI solutions for enterprise clients.

Responsibilities:

• Design and implement end-to-end AI solutions on the Microsoft Foundry platform, leveraging the model catalog (11,000+ models including OpenAI GPT, Anthropic Claude, Meta Llama, Mistral, and open-source options) to select, benchmark, and deploy the optimal models for each client use case
• Build and orchestrate intelligent agents using Foundry Agent Service, including single-agent flows (Foundry Classic), multi-agent workflows with visual orchestration, and hosted agent deployments using frameworks such as LangGraph, Semantic Kernel, and AutoGen
• Implement Retrieval-Augmented Generation (RAG) architectures using Foundry IQ (powered by Azure AI Search), including knowledge base configuration, grounding pipelines, citation-backed responses, and agentic RAG patterns with user access permission enforcement
• Fine-tune foundation models for domain-specific tasks using Foundry's fine-tuning capabilities, including dataset preparation, hyperparameter optimization, training pipeline configuration, and evaluation of fine-tuned model performance against baseline benchmarks
• Design and execute evaluation pipelines using Foundry's built-in evaluation tools, including automated quality loops (dataset creation, evaluation creation, evaluation comparison), red teaming exercises, and behavioral alignment assessments
• Integrate AI solutions with enterprise systems using Foundry Tools, MCP (Model Context Protocol) servers, and Azure Logic Apps connectors to enable seamless connectivity to platforms such as SAP, Salesforce, Dynamics 365, and custom APIs
• Implement AI governance and observability practices using Foundry Control Plane, Microsoft Defender integration, Entra ID for identity and access management, and Azure Policy for fleet-wide enforcement
• Leverage Foundry's memory capabilities to build agents with long-term context retention, automatic extraction and consolidation, and tailored user experiences across sessions
• Collaborate with cross-functional teams to translate business requirements into technical AI architectures, produce detailed design documentation, and guide clients from proof-of-concept through production deployment
• Participate in Agile ceremonies including sprint planning, standups, and retrospectives with the delivery team

Requirements:

• 7+ years of experience in machine learning, deep learning, or AI engineering, with at least 2 years of hands-on experience on Azure AI Foundry or Microsoft Foundry
• Strong theoretical and applied expertise in ML/DL fundamentals: neural network architectures (transformers, CNNs, RNNs, GANs), training methodologies, optimization algorithms, regularization techniques, and model evaluation metrics
• Deep understanding of large language models, including transformer architecture, attention mechanisms, tokenization, prompt engineering, fine-tuning (LoRA, QLoRA, full fine-tuning), and inference optimization (quantization, distillation, speculative decoding)
• Production experience deploying models at scale using Azure ML managed endpoints, Foundry model deployments, and container-based serving infrastructure (AKS, Container Apps)
• Proficiency in Python with strong experience in ML frameworks (PyTorch, TensorFlow/Keras) and AI orchestration libraries (LangChain, LangGraph, Semantic Kernel)
• Hands-on experience with RAG architectures, including vector databases, embedding models, chunking strategies, hybrid search (keyword + semantic), and reranking pipelines
• Solid understanding of MLOps practices: CI/CD for ML pipelines, model versioning, experiment tracking, automated retraining, and drift monitoring
• Experience implementing responsible AI practices, including fairness assessments, content filtering, toxicity detection, and model transparency reporting
• Microsoft certifications: Azure AI Engineer Associate (AI-102), Azure Data Scientist Associate (DP-100), or Azure Solutions Architect Expert (AZ-305)
• Experience with Foundry's latest capabilities: Agent-to-Agent (A2A) protocol, Foundry MCP Server, Model Router, and the unified azure-ai-projects SDK (v2)
• Background in specific industry AI applications such as financial services, healthcare, manufacturing, or energy
• Experience with multi-modal AI systems (vision, speech, text) and generative media models
• Contributions to open-source ML/AI projects or published research in ML/AI conferences or journals
• Familiarity with competitive AI platforms (AWS SageMaker, Google Vertex AI) for comparative solutioning