Architected a 2-phase analysis pipeline on AWS Bedrock with 3D classification, LLM orchestration, and 99.8% context compression, reducing data from 64 MB to 30 KB

Built an agentic AI chatbot with 4-tool orchestration, real-time FastAPI/SSE streaming, and context management within 50K-token budget

Designed Click-to-Compute notebook platform on ECS Fargate with async API provisioning, JWT proxy, Redis sessions, and per-user EFS isolation

Led a team of 6 data professionals delivering GenAI and ML solutions across AdTech clients with full lifecycle ownership from experimentation to production MLOps

Built a LangGraph-based multi-agent system enabling natural language exploration of large datasets via RAG with re-rankers, Vespa vector search, and schema-aware SQL generation

Launched GenAI-based persona summarization tool reducing manual efforts by 90% and deployed ML models serving 256 million users with 83% precision

Architected a scalable Next Best Offer recommendation system optimizing customer responses across targeted campaigns for 30M+ customer base

Engineered feature engineering pipelines with automated transformations, achieving 15% increase in model accuracy across analytics initiatives

Designed a data-driven customer engagement engine leveraging ML models and built an Automated Feature Shift Detection tool for production monitoring

Built forecasting pipelines for bakery chain using Python, PySpark, and AWS, reducing data processing time by 40% through Spark-based optimizations

Optimized Spark-based ETL solutions, cutting processing latency and improving end-to-end pipeline throughput

Architected data pipelines for seamless data transfer between multiple sources while ensuring data integrity and accessibility

Led migration of HR warehousing platform to big data architecture, achieving 30% reduction in data processing costs

Built PySpark-based analytics solutions with Tableau integration for HR management insights and delivered comprehensive analytics solutions using Hadoop ecosystem