AI Platform Architect

Role

As AI Platform Architect, you own the AI and ML layer of the Databricks platforms Visium builds for clients on Azure: feature pipelines, model lifecycle, serving, monitoring, governance, and the templates AI Engineers use to reach production. You build the platform underneath AI use cases, not the use cases themselves, so that the solutions built on it remain reliable, scalable, secure, and trustworthy in production.

You combine hands-on ML platform engineering with solution architecture, bringing the experience of having designed and built such platforms before. You shape the architecture with client leadership, assess workloads, identify bottlenecks, and turn technical requirements into a target design and an actionable delivery proposal. You work closely with AI Solutions Architects, AI and ML Engineers, Data Engineers, and platform teams.

Responsibilities:

Design, stand up and operate ML platforms on Azure Databricks, including feature pipelines and feature management, MLflow experiment tracking, model registry, training pipelines, and batch or online serving. Define and implement model lifecycle standards covering versioning, reproducibility, evaluation gates, promotion, deployment, rollback, and retraining. Build production-ready deployment patterns, including models exposed through APIs, automated MLflow workflows, and CI/CD pipelines for models and data. Own production monitoring for ML systems, including data and concept drift, model quality, latency, reliability, and cost, with appropriate alerting and retraining triggers. Design scalable Databricks workloads, identify architectural and performance bottlenecks, and translate findings into clear solution designs and delivery proposals. Lead decisions on ML tooling, integration patterns, and platform evolution, with Databricks as the primary platform and Azure ML used where appropriate. Build reusable templates, reference implementations, and accelerators that enable AI Engineers to move solutions into production independently. Establish ML governance practices, including lineage, audit trails, model documentation, access controls, and compliance-ready operating processes. Support GenAI delivery when needed, including deployment and monitoring of RAG and LLM applications using Azure OpenAI or Azure AI Foundry, vector search, evaluation, and guardrails. Apply the client’s existing cloud, security, networking, and operating standards to ML systems and integrate with the established landing zone and data platform. Ensure ML platforms and workloads are reliable, cost-efficient, secure, maintainable, and observable. Act as lead architect on client projects and the main client-facing contact for technical platform questions, discussing architecture with client CTOs and technical leadership. Contribute to the technical part of proposals and RFPs, and shape Visium’s strategy for designing, deploying and managing AI and ML platforms. Partner with engineering teams and senior stakeholders to guide implementation, mentor contributors, and accelerate adoption, helping shape a growing team as one of its senior technical anchors.

Requirements

To succeed in this role, you bring a combination of expertise, experience, and skills including:

5+ years of experience in MLOps, ML platform engineering, or a similar role, typically building on earlier experience in data science or ML engineering, with significant hands-on experience running ML systems in production. Proven experience designing and standing up an ML or AI platform yourself, not only building on one designed by others. Proven experience taking models to production and keeping them reliable, with concrete examples covering deployment, monitoring, incident resolution, and lifecycle management. Deep hands-on expertise in Databricks, especially MLflow, Unity Catalog, Model Serving and Asset Bundles. Azure Databricks is strongly preferred. Experience with Azure ML is a strong advantage. Strong experience designing Databricks solutions, diagnosing workload bottlenecks, and producing clear technical designs and implementation proposals. Software-engineering skills for MLOps and platform work, including packaging, testing, APIs, and maintainable automation. Deep experience building CI/CD pipelines for models and data, beyond infrastructure and application deployment alone. Strong knowledge of containers and model-serving patterns. Experience with Kubernetes or AKS is beneficial but not required. Confidence with Terraform and Infrastructure as Code, particularly when integrating ML services into an existing Azure landing zone. Working knowledge of cloud networking and security, including IAM, encryption, secrets management, governance, and observability.