⛯
Rathin Sharma★ Senior Analytics Engineer @ Cooke Inc.
I ✦ woo data into decisions
Data engineering that drives business outcomes.
I’m a Data & Analytics Engineer who turns messy data into decision-ready insight. I work with Snowflake, Airflow, dbt, Power BI, and thoughtful modeling that business users actually adopt.
I treat data like a product. Contracts, lineage, and quality get the same rigor as code. I challenge assumptions and hunt for the simple design beneath complexity.
Outside of work, I explore ideas across tech and economics, and refine them until I can explain them simply.
Where I’ve Left a Mark
A clear line from inputs to outcomes. Showing the levers pulled and the measurable results they created
Inventory Aging
- Built a unified item–vendor data model across 10+ ERPs
- Introduced a decision layer to identify and flag “dead inventory”
- Partnered with sales to prioritize selling older stock first
- Created materialized marts with contract tests to ensure accuracy
Result: ~60% reduction in dead inventory and significant warehousing cost savings
Quality Inspection
- QC images centralized in Azure Blob storage
- Models exposed as image URLs and visualized them
- Ops PowerBI dashboard with drillable defects
- Increased visibility of defects across the division
Result: ~30% faster inspection cycles and clearer root causes
Service Desk Response
- Capability mapping with a bridging mechanism across 12 teams
- Streamlined weekly Change Management process
- Automated snapshots in dbt to maintain history
- Improved workload visibility for proactive staffing
Result: ~15% reduction in response time and better staffing
Building Warehouse Management System
- Designed end-to-end data architecture to integrate ERP, IoT, and inventory modules
- Implemented real-time stock visibility and location tracking
- Developed automated replenishment and picking logic
- Integrated reporting layer with KPIs for throughput, accuracy, and turnaround time
Result: ~25% faster order fulfillment, reduced stockouts, and improved warehouse efficiency
Canonical Models for Yield
- Unified yield data definitions across multiple plants and ERPs
- Standardized transformation logic in dbt for consistent metrics
- Automated historical snapshots to track yield trends
- Exposed models to BI tools for plant-level and enterprise-wide analysis
Result: ~20% improvement in yield reporting accuracy and faster operational decision-making
Procurement Cost Saving
- Consolidated purchasing data across 10+ ERPs for spend visibility
- Identified price discrepancies and vendor performance gaps
- Trigger negotiations for better contracts based on data-driven insights
- Implemented tracking to ensure sustained cost savings
Result: ~12% reduction in procurement costs and improved vendor terms
The Data Lifecycle
I work end-to-end. Click bars to see how.
AI / ML
Testing / Experimentation
Reporting / Business Intelligence
Data Transformation / Stewardship
Storage / Data Engineering
Collection / Source Systems
Collection / Source Systems
ERPs, apps, IoT via Fivetran/Boomi/custom connectors/Azure Functions. CDC/events designed with data contracts from the start.
Technologies I use
Depth over logos. What I run in practice.
Ingest & Store • Model & Transform • Serve & Observe
Ingest & Store
FivetranAirbyteKafkaEvent HubS3/BlobSnowflake
- CDC and batch connectors with cost guardrails
- Staging schemas with freshness tests
- Multi-zone storage and retention strategy
Model & Transform
dbtSQLPythonGreat ExpectationsData Contracts
- Medallion + Kimball hybrid
- Contracts, lineage, and CI checks on PR
- Snapshotting and slowly changing dimensions
Serve & Observe
Power BIMetrics LayerAPIsdbt SemanticAirflow
- Metric definitions shared across tools
- Drillable dashboards with SLAs
- Observability and cost tracking
What informs my approach
Designing Data-Intensive Applications by Martin Kleppmann
A deep dive into building reliable, scalable, maintainable data systems. It shapes how I design ingestion, storage, and serving so insights are trustworthy and timely.
Fundamentals of Data Engineering by Joe Reis & Matt Housley
Pragmatic trade-offs across ingestion, modeling, orchestration, and consumption. Clarifies ETL vs ELT choices and how to build systems that last beyond the first quarter.
The Making of a Manager by Julie Zhuo
Tactics for setting expectations, giving feedback, and building trust. Guides how I mentor engineers and run projects without surprises.
A Fun Activity : ETL vs ELT
See what I do, like a 10 year old 😉 → cleaned data through stages. ETL = Transform before Load. ELT = Load before Transform.
Choose flow and run
InteractiveExtract → Transform → Load
1. Extract
id
item
price
1
nitrile gloves
13$
2
NITRILE GLOVES
1.5 $
3
Gloves, nitrile
$7
4
nitrile glove
$12.00
5
gloves-nitrile
USD 5
6
nitrilegloves
02$
7
Glove: Nitrile
10.5$
2. Transform
id
item
price
1
Nitrile Gloves
$13.00
2
Nitrile Gloves
$1.50
3
Nitrile Gloves
$7.00
4
Nitrile Gloves
$12.00
5
Nitrile Gloves
$5.00
6
Nitrile Gloves
$2.00
7
Nitrile Gloves
$10.50
3. Load
id
item
price
1
Nitrile Gloves
$13.00
2
Nitrile Gloves
$1.50
3
Nitrile Gloves
$7.00
4
Nitrile Gloves
$12.00
5
Nitrile Gloves
$5.00
6
Nitrile Gloves
$2.00
7
Nitrile Gloves
$10.50
ETL = transform before load. ELT = load then transform in-warehouse.
Let’s talk data
Have a hairy problem or an idea to validate? I can help you get to decisions faster.