Industry DepthRetail & CPGEnergy & ResourcesFinancial ServicesHealthcareSaaS & TechnologyManufacturing
Focus Areas
Enterprise AI & Interoperability
Built Snowflake's Interoperability offering — 5 WORM patterns for Snowflake + Databricks coexistence: External-First CLD, SF-Managed Iceberg, CDC Stream Sync, CLD + Cortex AI, Delta Direct
5-part technical series in Snowflake Builders Blog · SHERPA Iceberg Playbook (16 modules) · WAF Open & Interoperable Lens (25 additions) · Cortex Code skill v5 (44+ downloads)
Apache Iceberg Compatibility Matrix: merged PR adding Snowflake Horizon Catalog across 19 platforms at icebergmatrix.org
AI production wins across Financial Services, Healthcare, Technology, and Philanthropy verticals — Agentic Finance AI, DocAI, Cortex Analyst, and Interoperability deployments
Blueprint Manager co-creator (Snowflake-Labs, announced Summit 2026); Iceberg SME
Financial Services: Fortune 50 banking and payments (YugaByte SE leadership)
Healthcare & AI: Production AI deployments — DocAI, Cortex Analyst, Agentic Finance AI
About
Most SE leaders came up the SE track. I spent a decade as the enterprise customer first — building petabyte-scale data platforms at Sherwin-Williams and Weatherford — before leading SE and SA teams at YugaByte and Snowflake. I understand what buyers actually care about because I’ve been the buyer. That loop changes how I build teams, run technical deals, and design go-to-market.
Career
Snowflake · Sr. Manager, Solutions Architecture — US West
Head of Solutions Architecture, Enterprise AI & Interoperability
Oct 2025 – Present
Leads a Solutions Architecture team across enterprise accounts spanning Financial Services, Healthcare, Retail & CPG, and Technology verticals
Originated and built Snowflake's Interoperability offering — a repeatable WORM architecture framework enabling Snowflake + Databricks coexistence without migration or data duplication; 5 validated patterns covering External-First, SF-Managed Iceberg, CDC Sync, CLD + Cortex AI, and Delta Direct
Technical thought leadership: co-built Blueprint Manager (Snowflake-Labs, announced Summit 2026); Iceberg SME; authored SHERPA Iceberg Playbook; contributed to Snowflake's Well-Architected Framework; published on Snowflake Solutions + Labs; released Cortex Code skill for field self-service
Published 5-part Snowflake–Databricks Interoperability series in Snowflake Builders Blog; merged community OSS contribution to icebergmatrix.org — added Snowflake Horizon Catalog across 19 platforms
AI production deployments across Financial Services, Healthcare, Technology, and Philanthropy — Agentic Finance AI, Cortex Agent, DocAI, Cortex Analyst, and Interoperability in production
Weatherford International · Sr. Director, Enterprise Data & Architecture
Head of Enterprise Data Platform & Data Governance
2024 – 2025
Owned AI-ready data infrastructure with $23M P&L; petabyte-scale operations for 300+ global developers across 5 SaaS energy products
Architected Centro™ (real-time drilling analytics), CygNet® SCADA (enterprise monitoring), ForeSite® (AI/ML production optimization), and Unified Data Model — multi-cloud: AWS, Azure, IoT/NVIDIA Jetson
The complete software suite architected during this tenure — Centro™, CygNet® SCADA, ForeSite®, and Unified Data Model — is central to Weatherford's Strategic Collaboration Agreement with AWS (May 2025); drove the data architecture foundation for the partnership and contributed to SCA negotiations
Built the company's first Enterprise Data Governance & Data Quality Office: 100+ stakeholder interviews, 3–4 core policies ratified at Executive Council level
YugaByte DB · Director of Sales Engineering — North Americas
Head of Sales Engineering, North America
2021 – 2024
Led NA SA team against $56M annual quota targeting Fortune 50 accounts across Financial Services, Retail, Media, and Manufacturing
Built and scaled the SE team from ground up; designed GTM motions across AWS, Azure, and GCP; grew partner revenue to $2M+ annually
Represented YugaByte at Snowflake Summit, Money 20/20, AWS re:Invent, Gartner D&A, and KubeCon
The Sherwin-Williams Company · Global Head of Data Engineering
VP-Level: Data Engineering, DaaS, Retail Analytics
2011 – 2021
Built and led 70+ data engineers across 5 countries; Data-as-a-Service platform serving 60K+ users, supporting $4B+ quarterly retail operations
1,300+ data pipelines processing 1PB+ daily; scaled Snowflake to 3,000+ Tableau users and 40+ data marts in 24 months
Architected consumer data flywheels, smart pipeline tooling, and automated data quality frameworks used across the $18.4B enterprise
Oracle Corp · Hitachi Consulting · Sierra Atlantic
Lead ETL Developer / Software Engineer
2006 – 2011
Data engineering and application delivery for Fortune 100 enterprises; ERP manufacturing, supply chain, and financial data pipelines across Latin America and North America
Core Skills
SE Team LeadershipSales Engineering ManagementAI/ML ArchitectureAgentic AI (Cortex)Snowflake InteroperabilityApache IcebergDatabricks CoexistenceEnterprise Account StrategyGTM Offering DevelopmentData GovernanceDistributed Systems / DBaaSTechnical Enablement at ScaleRetail & CPGEnergy & ResourcesFinancial Services
Education
MBA, Finance Cleveland State University · 2015
Executive Leadership Certificate Case Western Reserve University · 2017
Recognition
❄ Snowflake Data Super-Hero 2020–2023
SIM — Midwest RLF Class of 2020
Cleveland Leadership Center Civic Institute 2016
Boards & Civic
Esperanza Inc. — Board Chair, Governance 2017–2024
A 5-part technical series published in the Snowflake Builders Blog covering the architecture, patterns, and AI integration for running Snowflake and Databricks together — without migration, without duplication. Built from real-world enterprise deployments across Financial Services, Philanthropy, Retail, and Technology verticals.
What's Most Important When Building New Data Products?2021 · Fivetran / Hashmap Webinar
The Importance of Data Engineering2021 · The Analytic Mind
Managing Data at Global Scale2021 · Behind the Data Cloud · Snowflake
Conference Circuit
Snowflake SummitGartner Data & AnalyticsMIT Sloan Data SymposiumOracle Open WorldMoney20/20Dell Tech WorldAWS re:InventKubeConPostgreSQL ConferenceCDO & Data Leaders SummitTechEquity AI Forum
Building an Enterprise Data Governance & Data Quality Office
How I built the first enterprise-wide DG&DQ program at Weatherford — 100+ stakeholder interviews, NIST/GDPR/ISO alignment, Executive Council ratification.
Weatherford International, 2024
Blog · Feb 2024
Azure Storage Performance: Deep Dive with Distributed Database Workloads
Benchmarking Azure Premium SSD, Premium SSD v2, and Ultra SSD under TPC-C workloads with YugabyteDB — practical guidance on disk type selection.
Published Feb 1, 2024
Blog · Nov 2023
The Future of Money and Financial Services
Five takeaways from Money 20/20 USA — AI ethics in banking, quantum computing's threat to encryption, embedded finance, and the balance of automation vs. human oversight.
Open to conversations about enterprise AI architecture, Snowflake Interoperability, building high-performing SE teams, or speaking opportunities across Retail, CPG, Energy, and Finance.
Build data platforms where AI participates in the full lifecycle — generating code,
writing documentation, triaging failures — while humans stay in charge of intent,
definitions, and accountability. From context pipelines to agentic orchestration
to production evals, grounded in a single running case study built on Snowflake.
Snowflake & CortexAnthropic Claude APIdbt · Airflow · MCPGraphRAG · OpenLineageEU AI Act & Governance
Most data engineering books stop at moving bytes. This one starts where the hard problems actually live:
coordinating meaning, managing change, and maintaining trust as the business evolves faster than the
documentation can keep up — and AI agents are now consuming your data at scale.
AI-Based Data Engineering builds a complete platform from the ground up — context pipelines
that give models what they need to reason correctly, structured generation patterns that constrain what
models can produce, and evaluation frameworks that catch failures before they reach production.
Every pattern is grounded in OpsPulse, a single running case study that threads through
all 14 chapters so you see every concept applied to the same real system.
The platform starts at an AI-readiness score of 4 out of 18 and ends at
18 out of 18 — one dimension at a time.
SQL Generation, Guardrails, and Self-Healing Pipelines
16/18 SQLGuardrail · eval loop
Chapter 10
Orchestrating AI Pipelines with Airflow
16/18 deferrable operators
Part 4 — Governance and Scaling
Chapter 11
Evals and AI Observability in Production
17/18 RAGAs · DeepEval · OTel
Chapter 12
Governance and Security for Context and Agents
18/18 OWASP LLM · EU AI Act
Chapter 13
Scaling the AI Data Platform
18/18 team topology · adoption
Chapter 14
Reference Architecture and What Comes Next
18/18 synthesis · roadmap
From the Book — The OpsPulse Case Study
"OpsPulse starts in a realistic initial state: incomplete documentation, divergent metric
definitions, no golden datasets, and an AI-readiness score of 4 out of 18. By the end of
the book, you will have built the context infrastructure, agentic workflows, and evaluation
systems that turn it into a production-grade AI data platform."
OpsPulse is a global operational analytics initiative combining ERP, CRM, IoT telemetry,
and support signals. Four teams each have their own definition of "active customer" —
producing four different numbers from the same source data:
Sales ops → 14,230 (contract value > 0)
Product analytics → 9,847 (login in last 30d)
Customer success → 11,502 (no churn-risk flag)
Finance → 8,319 (recognized revenue this quarter)
Every chapter of the book resolves one more dimension of this problem — until the platform
can answer "what is our current active customer count?" with a single, traceable, reproducible number.
— Chapter 1: What Is AI-Based Data Engineering? · AI-Based Data Engineering, Packt 2026
Richie spent a decade as the enterprise customer — building petabyte-scale data platforms at
Sherwin-Williams and Weatherford International — before moving to SE and SA leadership at
YugaByte and Snowflake. He built Snowflake's Interoperability offering, co-created Blueprint Manager,
and has led AI production deployments across Financial Services, Healthcare, and Retail. He writes at
the Snowflake Builders Blog and speaks on enterprise AI architecture and data platform design.
Six dimensions. Eighteen points. One arc from raw data platform to production-grade AI infrastructure.
Every point is a concrete, measurable property of your data estate — not a maturity label.
Data platform can support human workflows but AI agents will fail silently — wrong answers, unauditable outputs, no quality gates.
7 – 12
Targeted Investments
Selective AI use cases are viable. Specific dimensions are blocking production; targeted fixes unlock the next capability tier.
13 – 18
Production AI Ready
AI agents can reason correctly, produce auditable outputs, operate within governance boundaries, and be evaluated for correctness.
The 6 Dimensions
Each dimension is scored 0–3 (None / Partial / Mostly / Complete). The six are independent — a weakness in any one undermines the others, because AI systems expose all six failure modes simultaneously.
D1
Schema Freshness
Column-level descriptions in INFORMATION_SCHEMA, kept current. The model cannot reason correctly about data it cannot describe.
OpsPulse end state: 3/3 · Addressed in Ch2, Ch8
D2
Semantic Coverage
MetricFlow semantic models and metric definitions. Without semantics, AI generates plausible-but-wrong answers to business questions.
OpsPulse end state: 3/3 · Addressed in Ch1, Ch4
D3
Test Coverage
dbt data_tests: or equivalent quality assertions. Quality must be a gate before AI inference, not a monitor after the fact.
OpsPulse end state: 3/3 · Addressed in Ch9, Ch11
D4
Eval Readiness
Golden datasets, snapshot history, regression harnesses. AI outputs cannot be verified without reference data to compare against.
OpsPulse end state: 3/3 · Addressed in Ch11
D5
Governance Scope
Column masking policies, row access policies, audit logging. AI agent queries must be governed with the same rigor as human queries.
OpsPulse end state: 3/3 · Addressed in Ch7, Ch12
D6
Lineage Coverage
Upstream and downstream lineage queryable at column level. Without lineage, AI outputs are unauditable and incidents are undiagnosable.
OpsPulse end state: 3/3 · Addressed in Ch2, Ch6
The Chapter-by-Chapter Arc
OpsPulse starts at 4/18 — a realistic baseline for most enterprise data platforms. Each chapter closes one or more dimension gaps. The score is the exact state of the platform at the end of that chapter.
4
4/18
Ch 1
baseline
8
~8/18
Ch 2
+4 catalog
9
~9/18
Ch 3
+1 patterns
11
11/18
Ch 4
+2 prompting
12
~12/18
Ch 5
+1 context
13
~13/18
Ch 6
+1 graph
14
~14/18
Ch 7
+1 MCP
15
15/18
Ch 8
+1 ingestion
16
16/18
Ch 9
+1 SQL gen
16
16/18
Ch 10
holds
17
17/18
Ch 11
+1 evals
18
18/18
Ch 12
+1 governance
18
18/18
Ch 13
holds
18
18/18
Ch 14
complete
Dimension Fill Map
Shows each dimension's score (0–3) at the end of each chapter. Darker blue = higher score. Scroll horizontally on small screens.
Dimension
Ch1
Ch2
Ch3
Ch4
Ch5
Ch6
Ch7
Ch8
Ch9
Ch10
Ch11
Ch12
Ch13
Ch14
D1 Schema Freshness
1
2
2
2
3
3
3
3
3
3
3
3
3
3
D2 Semantic Coverage
0
1
1
2
2
2
2
3
3
3
3
3
3
3
D3 Test Coverage
1
1
1
2
2
2
2
2
3
3
3
3
3
3
D4 Eval Readiness
1
1
2
2
2
2
2
2
2
2
3
3
3
3
D5 Governance Scope
1
1
1
1
1
1
2
2
2
2
2
3
3
3
D6 Lineage Coverage
0
2
2
2
2
3
3
3
3
3
3
3
3
3
Total
4
8
9
11
12
13
14
15
16
16
17
18
18
18
* Scores represent approximate OpsPulse state at chapter end. Intermediate chapters hold scores while building enabling infrastructure.
Score Your Platform
Select the level (0–3) that best describes your current data platform for each dimension. Your score updates live. This is the same 18-point checklist used in AI-Based Data Engineering Chapter 1.
D1
Schema Freshness
None
D2
Semantic Coverage
None
D3
Test Coverage
None
D4
Eval Readiness
None
D5
Governance Scope
None
D6
Lineage Coverage
None
0/ 18
Foundation Required
Your platform needs foundational work before AI agents can operate reliably. Start with D1 and D6 — schema and lineage are the fastest path to +4 points.
AI-Based Data Engineering — Packt 2026
Score your platform. Then read the book.
Each chapter closes specific dimension gaps in a deliberate order — the OpsPulse arc takes you from wherever you are to 18/18, one concrete pattern at a time.
