Senior Data Engineer | Cloud & Analytics Engineering | Building Enterprise Data Platforms & Intelligence Systems
- 🔭 Senior Data Engineer with 4+ years delivering ETL modernization, cloud analytics, and lakehouse platforms across Banking, Semiconductor, Aerospace, and Healthcare
- 🏗️ Architecting end-to-end pipelines — ingestion → distributed processing → warehouse → ML → BI
- 🌱 Specializing in Snowflake, PySpark, Azure/Databricks, workflow orchestration, and data quality engineering
- 🤖 Building predictive and analytical models for demand forecasting, defect correlation, and risk scoring
- 💬 Ask me about Python, SQL, PySpark, Snowflake, Power BI, ETL/ELT, Cloud Architecture, and Data Quality
- ⚡ Focused on measurable business impact — 75% faster reporting, 50% less manual validation, 35% lower processing overhead
| Project | What It Does | Core Stack | Headline Result |
|---|---|---|---|
| 🚨 Complaint Early-Warning System | Detects conduct-risk incidents at banks from 4.7M real CFPB complaints, backtested point-in-time | Kafka / Event Hubs · Delta Lake · PySpark · Snowflake · dbt · Airflow · Power BI | 8/11 incidents caught, 29 days ahead of monthly MIS |
| 🏥 FHIR Clinical Lakehouse | Privacy-preserving patient record linkage across organizations, no raw PII shared | PySpark · Delta Lake · Snowflake · Bloom Filters (PPRL) · Power BI | 97.5% match recall @ 100% precision |
| ⚙️ ETL Modernization Framework | LLM-augmented legacy Perl/SQL → Python translation with behavioral equivalence proof | Python · Claude API · sqlglot · Great Expectations · Airflow | 85.7% auto-translation, 99.63% equivalence |
| 🏦 Banking Regulatory Platform | Multi-source reconciliation and Basel III / RBI compliance reporting | Databricks · Delta Lake · ADLS Gen2 · Event Hubs · Power BI | T+1 → sub-30 min cycle, 98%+ break detection |
| Supply chain hub with AOG risk scoring and intermittent-demand forecasting | PySpark · Snowflake · XGBoost · Airflow | At-risk parts flagged ≥72 hrs early (median 30 days) | |
| 🔬 Semiconductor Yield Intelligence | BOM-aware defect correlation with automated root-cause ranking | scikit-learn · networkx · Snowflake · dbt · SAP RFC | Root cause in 34 sec vs 2–3 days |
Each repo has a full write-up — architecture, design decisions, and benchmark methodology.
Languages & Scripting
Data Engineering & Big Data
ETL/ELT development · Data warehousing · Data modelling (star schema, SCD) · Data lake & lakehouse · Distributed processing · Data migration
Orchestration & Workflow
Cloud
Analytics & BI
KPI dashboard development · Executive reporting · Drill-through & hierarchy design · Dimensional modelling
Data Science & ML
Time-series forecasting (MASE, intermittent demand) · Anomaly detection · Statistical testing · Feature engineering · Leak-proof cross-validation
Data Quality & Testing
Reconciliation & anomaly detection frameworks · Custom rule engines · Regression suite generation · Freshness & schema tests
DevOps & Delivery
Integration & Data Formats
Performance & Optimization
Senior Systems Engineer — Tata Consultancy Services, Bengaluru · Apr 2023 – Present
- Led Snowflake analytics and Power BI reporting for semiconductor manufacturing — ~30% faster dashboard turnaround
- Built Python automation for validation, reconciliation, and anomaly detection — ~50% less manual validation effort
- Modernized legacy Perl ETL into Python architectures — ~40% lower dependency management effort
- Designed Spark/PySpark pipelines for banking datasets — ~35% fewer large-scale processing bottlenecks
- Delivered Power BI reporting workflows — ~75% faster reporting turnaround
Systems Engineer — Tata Consultancy Services, Bengaluru · Dec 2021 – Mar 2023
- Built enterprise data integration for high-volume aerospace manufacturing datasets
- Developed Python ETL and Snowflake workflows for healthcare data — ~30% better processing performance
- Shipped 15+ reusable ETL components and orchestration solutions — up to 75% faster delivery timelines
| Project | Metric | Result |
|---|---|---|
| Complaint Early-Warning | Labelled incidents detected | 8 / 11 (median 9.5 days after onset) |
| Lead over monthly MIS | 29 days (median) | |
| FHIR Clinical Lakehouse | Match recall / precision | 97.5% / 100% |
| Data quality score | 98.9–99.7% | |
| ETL Modernization | Auto-translation rate | 85.7% |
| Defects reaching production | 0 | |
| Banking Platform | Reconciliation cycle | T+1 → sub-30 min |
| Break detection automation | 98%+ | |
| Aerospace MRO | AOG early warning | 100% flagged ≥72 hrs early |
| Forecast accuracy | MASE 0.86 vs baseline | |
| Semiconductor Yield | Root-cause shortlist | 34 sec (was 2–3 days) |
| Cost impact quantified | $2.48M per batch |
B.E. Mechanical Engineering — Anna University, Chennai · 2016–2020
I build data systems that solve real problems — preventing aircraft groundings, eliminating duplicate patient records, catching yield defects before they cascade into millions in losses. Every project balances engineering rigor (testing, gatekeeping, reproducibility) with pragmatic delivery. Data is only valuable if it moves someone to action, so I obsess over the last mile — the dashboard, the alert, the metric that changes a decision.
