Data and Growth Engineer (Data Function Lead) at EssentiallySports

Led the data and AI function for a sports media house generating ~30M impressions. Built production agentic AI systems and structured RAG pipelines using GCP Cloud Run, Gemini, and AWS Athena to generate Slack recommendations and reduce spam/promotions while improving newsletter CTR. Implemented a multi-agent Google ADK content auditor on Vertex AI/Gemini with tool use, prompt chaining, proprietary-playbook grounding, and gold-set evaluation to serve structured JSON scores to users. Built human-in-the-loop LLM enrichment pipelines storing validated features to AWS RDS and automating dbt incremental runs powering Apache Superset analytics. Architected ETL/ELT flows and governed warehouse models for external partnership data (~$500K), including probabilistic identity resolution for ~80M anonymous readers, enabling attribution and recommendation improvements. Drove growth metrics including scaling newsletter-attributed pageviews from 1.5M to 8.5M, doubling subscriber lifetime, and improvi

Data Scientist (contract) at Blitz Bangalore

Built performance dashboards for Logistics Ops and FinOps in Metabase and Tableau, tracking 4,000+ orders and 30+ daily KPIs. Optimized PostgreSQL databases and CTE-based SQL workflows orchestrated with Apache Airflow, reducing system-wide query runtimes by ~70%.

AI / ML Engineer & Consultant at Aisens.ai

Delivered end-to-end applied AI systems for international clients spanning multi-agent orchestration, LLM document intelligence, ML pipelines, API integration, and cloud deployment. Built a five-agent recruiting system on Vertex AI using Google ADK with MCP integrations for LinkedIn/Indeed, achieving ~20x faster shortlisting and recruiter acceptance rate above 95% via human-in-the-loop evaluation. Engineered a layout-aware multilingual document intelligence pipeline for structured extraction from furniture catalogs. Implemented a human-in-the-loop cross-marketplace taxonomy-mapping pipeline, price prediction with Random Forest, and reproducible training + SageMaker deployments. Developed web and document extraction for real-estate agents using Playwright/Selenium/BeautifulSoup/Tesseract with LLM cleanup and rule-based deduplication. Led production AWS log-analytics/data engineering work for HR/payroll logs with monitoring and ETL automation, reducing manual verification time by 80%.