Data Engineer & GenAI Developer
Building production data platforms and LLM-powered systems. 3 live projects, 20M+ records processed, 6,000+ users.
I ran MAS Nexus, my own engineering practice, for over two years: 27 months as an embedded data engineer for a client on vendor basis (data pipelines, reporting layer, tooling decisions, AI features), plus the platforms below built alongside. Reference letter available on request.
Based in Germany as a Chancenkarte (Opportunity Card) holder, authorized to work anywhere in Germany and available immediately.
Serverless ELT pipeline over 20M+ geopolitical news events from GDELT, in continuous scheduled operation for 8+ months.
- 6,000+ unique visitors in the first few months, 100+ new users daily, zero advertising
- 100K+ events ingested daily on an hourly refresh cycle
- Dual AI chat: text-to-SQL and RAG over vector embeddings
- Slowest stage rewritten in Polars (~10x faster than Pandas)
- $0/month: MotherDuck + Dagster + GitHub Actions on free tiers
Python Polars dbt MotherDuck Dagster LlamaIndex
Real-time streaming platform for AI-powered cart-abandonment recovery.
- Azure Event Hubs → Delta Live Tables, <500ms end-to-end latency
- LLM-personalized recovery messages across 5 customer archetypes
- Medallion lakehouse (Bronze/Silver/Gold) with DLT expectations
- A/B testing with z-scores at 95% confidence
Databricks Azure Event Hubs DLT Cerebras dbt Streamlit
Databricks lakehouse over 100K+ orders from the Brazilian Olist marketplace.
- Medallion architecture with Delta Lake
- Kimball star schema: fct_orders + customer/product/seller dimensions
- Unity Catalog governance with documented lineage
Databricks Delta Lake SQL Unity Catalog Power BI
Anki deck with 5,000+ German words across A1 to C1, built while learning German myself. Near 3,000 downloads, all organic.
- Designed the build pipeline with AI: word selection by frequency and learner need
- Neural audio for every word, images, verb conjugations
Python Anki Text-to-Speech
Interactive Excel dashboard analyzing 12,894 data job postings.
- 70+ step Power Query ETL, 228K+ rows in Power Pivot with custom DAX
- 4,000+ messy job titles standardized into role categories
Excel Power Query Power Pivot DAX
- Data engineering: Python · SQL · Databricks · Delta Lake · Azure · DuckDB/MotherDuck
- Pipelines: dbt Core · Dagster · Polars · GitHub Actions
- AI/LLM: RAG · LlamaIndex · Cerebras · Voyage AI embeddings
- Visualization: Streamlit · Plotly · Power BI
Data Engineer & GenAI Developer. Open to Data Engineer, Analytics Engineer, AI Engineer, and Data Analyst roles in Germany. Available immediately, authorized to work in Germany (Chancenkarte, no sponsorship needed), English C1, German A2 (working toward B1).
B.Tech, Electrical & Electronics Engineering, SRM University, Chennai