At GlobalLogic on Google GCAS, I build and productionize AI/ML and GenAI solutions for advertising across data pipelines, model development, deployment, monitoring, and MLOps.
I developed AI Search Scout, a large-scale search trend forecasting pipeline using Transformer embeddings, semantic and vector search, clustering, feature engineering, and time-series forecasting to support multi-million-dollar advertising budget decisions.
I also built a RAG and LLM advertising intelligence platform with Gemini, hybrid retrieval, re-ranking, KPI computation, and grounding guardrails, alongside an agentic system for planning, tool calling, retrieval, and analytics workflows.
My research at CAIR-DRDO included PPO-based reinforcement learning agents, behavior cloning, and GAN-based data augmentation that improved task success by 20–25% and search completion by 15–18% over baseline models.
