Data & AI Leader · 15+ Years
Sumit Gupta
I design enterprise data architectures, build and scale the teams that run them, and ship AI products and accelerators that shorten delivery and cut cost. Lately: open-source tooling that makes data systems legible.
What I do
Architecture
- Enterprise-grade data architectures across 12+ systems
- End-to-end Databricks data platform, designed and delivered
- Snowflake Medallion architecture — integration framework, reporting layer, AI agents
- X-flake, a migration utility for large-scale Snowflake adoption
- Automation frameworks for CI/CD, deployment and monitoring
Leadership
- Scaled a programme team from 2 to 35 people
- Established Dev, Testing and Operations functions from scratch
- Architecture standards and governance models for Snowflake
- Hiring across projects; mentoring and learning paths for AI, Snowflake, AWS
- Reporting into CDO, CTO and client Data Heads
Products & accelerators
- Mini Beast Data Tool and Travel Nexus
- RAG-based agents for intelligent data processing
- Reusable data-engineering components for faster delivery
- Winner, Capgemini Data Scientist Hackathon
- Multiple Data Science and Cloud hackathon wins
Open source
Lineage, relationships, profiling, dimensional modelling and an AI-ready knowledge graph — built deterministically from your warehouse, dbt project and SQL. No LLM in the loop.
What breaks if I merge this? Takes a pull-request diff, walks the dependency graph, and reports blast radius, risk, owners and a test plan — with an exit code you can gate CI on.
Composable agents for data and backend engineering. Permission-gated tools, prompt-injection containment and an audit trail on every run. Runs on Bedrock, Anthropic or OpenAI.
Tech
| Warehouse | Snowflake / Redshift |
|---|---|
| Languages | Python / Scala / PySpark / SQL |
| Cloud | AWS / Azure / Databricks |
| Databases | PostgreSQL / MySQL / MongoDB / HBase / Neo4j |
| Data tooling | dbt / Spark / Fivetran / HVR / Workato / APIs |
| AI | AI agents / RAG / ML algorithms / NLP |