This is an individual-contributor technical leadership role reporting to the Director of Data Platform Engineering, with no direct reports initially. It is a platform-engineering role—not an application data-engineering, data-science, or MLOps role. You will engineer the shared foundation, guardrails, automation, observability, and self-service capabilities used by Data Engineering, Analytics, ML Engineering, AI Engineering, and MLOps.
Key responsibilities
Databricks platform ownership
Own Databricks account- and workspace-level capabilities across development, test, and production environments.
Automate workspace provisioning and configuration, including identity federation, SSO/SCIM integration, service principals, workspace assignments, entitlements, and administrative controls.
Administer Unity Catalog, including metastores, workspace bindings, storage credentials, external locations, lineage, auditing, and implementation of the access model defined by Data Governance and data owners.
Define approved patterns for serverless and classic compute, jobs, interactive compute, SQL warehouses, and specialized compute when required.
Maintain compute policies, runtime and library standards, tagging, auto-termination, quotas, compatibility testing, and LTS/deprecation plans that balance security, performance, reliability, and cost.
Infrastructure as Code (IaC), AWS integration, and security
Own the IaC strategy and implementation for the Databricks platform, building and maintaining reusable Terraform modules with the Databricks and AWS providers.
Manage supported account, workspace, Unity Catalog, compute, and AWS integration configurations through code, with environment isolation, automated plans and tests, approval controls, and configuration-drift detection and remediation.
Build tested CI/CD workflows using account, workspace, Unity Catalog, and shared-platform changes, including peer review, approvals, controlled promotion, and rollback.
Provide reusable Declarative Automation Bundle templates—formerly Databricks Asset Bundles—and secure deployment patterns. Workload teams remain responsible for their code and releases.
Consume enterprise AWS guardrails and partner with Cloud Platform to automate Databricks-specific networking, private connectivity, IAM, S3, KMS, DNS, secrets management, and connectivity with services such as Aurora and Kafka.
Implement approved controls for PHI and other sensitive data, including encryption, network isolation and egress, privileged access, logging, approval workflows, access reviews, and break-glass procedures.
Ensure privileged and account-level changes are reviewed and approved by designated Cloud Platform, Cybersecurity, or Data Governance partners based on the affected control.
Support security reviews, audit evidence, risk remediation, and production change-management requirements.
Reliability, observability, and FinOps
Establish platform service levels, health indicators, alerts, dashboards, escalation paths, and runbooks using Databricks system tables, audit logs, CloudWatch
Participate in a shared on-call and escalation rotation; lead diagnosis, root-cause analysis, and preventive action for complex Databricks and AWS integration incidents.
Implement cost attribution, budgets, tagging, showback reporting, and optimization recommendations across teams and workloads.
Track Databricks consumption against commercial commitments, forecast capacity, and provide technical analysis for planning and renewal decisions.
Own Databricks recovery patterns and testing in partnership with Cloud Platform and workload owners, aligned with agreed recovery objectives.
Data and AI platform enablement
Build paved-road templates, standards, documentation, and support mechanisms that increase self-service and reduce one-off administrative requests.
Partner with Data Engineering, Analytics, ML Engineering, AI Engineering, and MLOps to translate workload requirements into supported platform capabilities.
Configure shared MLflow and Models in Unity Catalog capabilities and enable approved training, inference, and model-serving patterns as they are adopted.
Troubleshoot platform issues without assuming ownership of pipeline logic, model development, evaluations, AI application behavior, or workload releases.
Required qualifications
Demonstrated ownership of Databricks on AWS in a multi-team production environment, including account- and workspace-level administration.
Strong experience with Unity Catalog, identity federation, service principals, compute policies, secure data-access patterns, lineage, and auditability.
Deep production experience with Terraform and IaC practices, including modular infrastructure built with the Databricks and AWS providers, state management, source control, automated testing, and CI/CD.
Strong AWS knowledge across IAM, VPC networking, private connectivity, S3, KMS, security groups, logging, and cross-account access.
Experience operating cloud data platforms containing regulated, confidential, or highly sensitive data.
Experience with observability, incident response, change management, capacity planning, and cloud cost optimization.
Experience leading a significant platform migration, workspace or metastore consolidation, runtime upgrade, or comparable modernization under production load.
Working knowledge of Spark and distributed computing sufficient to diagnose platform and workload-performance issues.
Proficiency in Python for automation, plus practical shell scripting and REST API integration.
Strong communication and influence skills, with demonstrated success establishing standards and self-service guardrails across organizational boundaries.
Preferred qualifications
Experience operating healthcare data platforms and implementing HIPAA-aligned controls for PHI.
Experience with MLflow, Models in Unity Catalog, model serving, GPU workloads, feature engineering platforms, or GenAI infrastructure.
Experience with multi-workspace, multi-account, or multi-region architectures and disaster-recovery testing.
Databricks and/or AWS professional certification. Certification is valuable but is not a substitute for production platform ownership.
Measures of success
Shared-platform configurations are reproducible, peer-reviewed, and version-controlled through infrastructure as code.
Privileged access and production changes follow least-privilege and auditable approval practices.
Within the first 90 days, establish measurable baselines for platform reliability, configuration drift, manual administrative effort, and unit cost; demonstrate sustained improvement against those baselines.
Teams deploy approved workloads through documented, reusable, self-service patterns.
Databricks cost and commitment utilization are visible, attributable, forecastable, and optimized without compromising agreed reliability or performance.
Ownership boundaries
This role owns the Databricks platform and implementation of approved technical controls. Cloud Platform owns enterprise AWS foundations and guardrails. Cybersecurity and Data Governance define policy; data owners approve business access. Data Engineering, Analytics, and AI/ML teams own their code, data products, models, evaluations, and releases. The ML/AI/LLM Operations function owns model promotion, deployment automation, production monitoring, and rollback.
ModMed Benefits Highlight: At ModMed, we believe it’s important to offer a competitive benefits package designed to meet the diverse needs of our growing workforce. Eligible Modernizers can enroll in a wide range of benefits:
United States
Comprehensive medical, dental, and vision benefits, including a company Health Savings Account contribution,
401(k): ModMed provides a matching contribution each payday of 50% of your contribution deferred on up to 6% of your compensation. After one year of employment with ModMed, 100% of any matching contribution you receive is yours to keep.
Generous Paid Time Off and Paid Parental Leave programs,
Company paid Life and Disability benefits, Flexible Spending Account, and Employee Assistance Programs,
Company-sponsored Business Resource & Special Interest Groups that provide engaged and supportive communities within ModMed,
Professional development opportunities, including tuition reimbursement programs and unlimited access to LinkedIn Learning,
Global presence and in-person collaboration opportunities; dog-friendly HQ (US), Hybrid office-based roles and remote availability for some roles,
Weekly catered breakfast and lunch, treadmill workstations, Zen, and wellness rooms within our BRIC headquarters.
PHISHING SCAM WARNING: ModMed is among several companies recently made aware of a phishing scam involving imposters posing as hiring managers recruiting via email, text and social media. The imposters are creating misleading email accounts, conducting remote "interviews," and making fake job offers in order to collect personal and financial information from unsuspecting individuals. Please be aware that no job offers will be made from ModMed without a formal interview process, and valid communications from our hiring team will come from our employees with a ModMed email address (first.lastname@modmed.com). Please check senders’ email addresses carefully. Additionally, ModMed will not ask you to purchase equipment or supplies as part of your onboarding process. If you are receiving communications as described above, please report them to the FTC website.
