I've built enterprise data platforms from the inside and sold them from the outside. I speak both languages fluently.
Before the resume, here's who you'd actually be working with.

Quota, revenue, and platform outcomes I can walk you through line by line.
| # | metric | value | source |
|---|---|---|---|
| 1 | Production quota, FY2018-22 | 0% | uspto |
| 2 | Pipeline throughput gain across production systems | 0% | cvs_health |
| 3 | Runtime reduction on enterprise cloud migration, zero data loss | 0% | cvs_health |
| 4 | Annual revenue growth contributed | 0% | findur_presales |
| 5 | Enterprise client stakeholders managed as technical liaison | 0+ | findur |
| 6 | Business units coordinated on a single enterprise migration | 0 | cvs_health |
| 7 | Years translating deep tech for lawyers, inventors & execs | 0 | uspto |
| 8 | Certifications: GCP Professional DE · Databricks Associate | 0 | verified |
I took an unusual path: I learned to communicate under pressure first, sell second, and build third. Most SEs go the other way, and it shows in my demos.
Every tool below has shipped something real: an enterprise migration, a live demo, or a weekend build.
| category | tools |
|---|---|
| cloud_and_data | GCPBigQueryDataprocDatabricksVertex AIAzure Data FactoryMicrosoft FabricTeradata |
| pipelines | Apache AirflowInformaticaETL / ELT |
| ai_and_analytics | AI/ML PlatformsPredictive ModelingLLM ToolingExecutive Dashboards |
| languages | PythonSQL |
| presales_craft | Technical DiscoverySolution ArchitectureProof of ConceptLive DemosExecutive Presentations |
| certifications | GCP Professional Data EngineerDatabricks Associate Data Engineer |
Messy inputs, clean outputs, nothing silently dropped. Discovery extracts what actually matters, architecture turns it into one clear answer, and a live proof on their own data is what loads the trust. Then it runs on a schedule: renewals.
QUERY PLAN --------------------------------------------- Deliver (renewals follow) -> Prove (live demo, their own data) -> Simplify (one clear answer) -> Discover (what matters) Scan on requirements (rows=50, loose_ends=0) Planning: one good discovery call Execution: faster than you'd think (1 row: trust)