Snowflake Administrator

AI Platform Enablement & Governance (core focus)

Cortex AI enablement: Enable, configure and govern Snowflake Cortex AI services — Cortex AISQL functions (AI_COMPLETE, AI_FILTER, AI_CLASSIFY, AI_AGG, AI_EXTRACT, embedding and similarity functions), Cortex Search, Cortex Analyst, Cortex Agents and Snowflake Intelligence — across development, test and production accounts.

AI access control: Design the RBAC model for AI consumption: which functional roles may call LLM functions, which models and regions are approved, cross-region inference settings, and privilege boundaries between AI builders, analysts and service accounts.

Model & data guardrails: Configure Cortex Guard and equivalent safety controls, restrict model allow-lists, and ensure masking policies, row-access policies and tag-based governance are enforced on data flowing into prompts and out of AI responses.

AI cost engineering: Instrument, forecast and control AI service consumption (credits per LLM function, per warehouse, per team) using resource monitors, budgets, account usage views and chargeback reporting. Establish alerting thresholds before overruns, not after.

Semantic layer operations: Support and operationalise semantic views and semantic models that power Cortex Analyst and natural-language querying — versioning, change control, accuracy regression checks and promotion across environments.

Unstructured & multimodal data: Administer stages, directory tables, file formats and document-processing (Document AI) pipelines that feed AI workloads, including retention and PII handling for unstructured content.

Snowpark & container workloads: Provision and govern Snowpark, Snowpark Container Services, external access integrations, network rules and secrets used by AI and ML applications.

AI observability & auditability: Build monitoring and audit trails for AI usage — who prompted what, against which data, using which model — sufficient to satisfy internal audit, client security reviews and regulatory scrutiny.

Responsible AI alignment: Translate clients' AI governance policies into enforceable Snowflake configuration, and document the control mapping for audits (ISO 27001, SOC 2, GDPR, HIPAA and client-specific frameworks as applicable).

2.2 Core Snowflake Platform Administration

Own account-level administration across multiple Snowflake accounts, editions, clouds and regions — including organisation accounts, replication, failover and business continuity.

Design and maintain a scalable RBAC framework using the access-role / functional-role / admin-role pattern to prevent privilege sprawl and keep access auditable.

Manage identity and authentication: SSO/SAML and SCIM integration with Microsoft Entra ID, key-pair authentication, mandatory MFA, service-account lifecycle and credential rotation.

Provision, size and tune virtual warehouses (including multi-cluster and Gen2 / adaptive compute), query acceleration, caching strategy and clustering keys.

Deliver measurable cost optimisation: warehouse right-sizing, auto-suspend policy, resource monitors, budgets, storage lifecycle, and elimination of idle or duplicated spend.

Tune query and workload performance using Query History, Query Profile, access history and ACCOUNT_USAGE / ORGANIZATION_USAGE views.

Administer data protection features: Time Travel, Fail-safe, zero-copy cloning, dynamic data masking, row-access policies, object tagging, network policies and private connectivity.

Operate Secure Data Sharing, reader accounts, listings and Snowflake Marketplace participation.

Manage governance tooling — Snowflake Horizon Catalog, Trust Center posture findings, lineage and data classification — and remediate findings on a defined cadence.

Automate platform operations as code: Terraform (Snowflake provider), Snowflake CLI, Python/SnowSQL scripting, Git-based object deployment, CI/CD and schema change management.

Support ingestion and integration: Snowpipe, Snowpipe Streaming, Openflow / connector-based ingestion, dbt, Azure Data Factory or Fabric, Kafka and third-party ELT tooling.

Own incident, change and problem management for the platform — RCA, runbooks, on-call participation and escalation handling with Snowflake Support.

Plan and validate release and behaviour-change adoption, feature-bundle testing and upgrade impact assessment.

2.3 Leadership, Advisory & Delivery

Act as senior escalation point and technical authority for Snowflake across delivery teams.

Mentor junior administrators and data engineers; run internal enablement sessions and build reusable admin accelerators.

Define and publish platform standards, naming conventions, security baselines and architecture decision records.

Contribute to pre-sales and client engagements: Snowflake assessments, migration and modernisation roadmaps, AI-readiness reviews, effort estimation and solution documentation.

Partner with client security, compliance and FinOps stakeholders during audits and quarterly platform reviews.

3. Required Skills & Experience

8+ years in data platform, database or cloud infrastructure engineering, with 4+ years of hands-on Snowflake administration at production scale.

Deep, demonstrable expertise in Snowflake RBAC design, security, governance and multi-account administration.

Hands-on production experience with Snowflake Cortex AI — AISQL functions, Cortex Search, Cortex Analyst or Cortex Agents — including enabling them securely and controlling their cost.

Strong SQL, including performance analysis via Query Profile and the ACCOUNT_USAGE / ORGANIZATION_USAGE schemas.

Practical grasp of LLM fundamentals relevant to platform ownership: token economics, embeddings and vector search, RAG architecture, context windows, grounding and hallucination risk, and model evaluation.

AI governance and data-privacy competence: PII detection and masking in AI pipelines, prompt and output logging, data-residency and cross-region inference constraints, model allow-listing.

Infrastructure-as-code and automation: Terraform, Python, Git, CI/CD for database and platform objects.

Working knowledge of at least one hyperscaler (Azure strongly preferred; AWS or Google Cloud Platform acceptable) — networking, private endpoints, identity and storage.

Cloud cost management / FinOps discipline with a track record of quantified savings.

Strong written communication: runbooks, standards documents, control mappings and client-facing reports.

4. Preferred / Differentiating Skills

SnowPro Advanced: Administrator certification; SnowPro Advanced: Architect or Data Engineer also valued.

Experience building or operating a production RAG or conversational-analytics application on Snowflake.

Snowpark Container Services, external access integrations and model-serving experience.

MLOps exposure: Snowflake ML, model registry, feature store, model monitoring and drift detection.

Experience with Snowflake Horizon Catalog, Trust Center, Openflow, dynamic tables and Iceberg tables.

Familiarity with Microsoft Fabric, Databricks or Azure Synapse for interoperability and comparative advisory work.

Regulated-industry delivery experience: HIPAA (healthcare), PCI-DSS, SEBI/RBI (BFSI) or GDPR-scoped engagements.

Prior consulting or client-facing delivery experience in a services organisation.

5. Certifications

Required: SnowPro Core Certification, or achieved within 90 days of joining.

Preferred: SnowPro Advanced: Administrator; Microsoft Azure Data Engineer Associate; Azure AI Engineer Associate (AI-102); a recognised AI governance credential such as IAPP AIGP.

6. How Success Is Measured (First 12 Months)

90 days: Full inventory and risk assessment of all Snowflake accounts; RBAC and AI-access gaps documented with a remediation plan.

6 months: Cortex AI usage governed by an approved, documented control framework with role-based access, model allow-lists and per-team cost visibility.

6 months: Measurable reduction in Snowflake compute spend without SLA degradation, evidenced through resource monitors and budget reporting.

12 months: Platform operations automated as code with CI/CD, zero critical audit findings, and a repeatable Snowflake + AI governance accelerator reusable across client engagements.