Stuti G.

About Stuti G.

I\'m a Machine Learning Engineer specializing in computer vision, natural language processing and production AI systems. Currently pursuing my M.S. in Data Science at the University of Washington, I build models that work in the real world, not just in notebooks.What sets me apart: I\'ve deployed AI across diverse domains - from biometric verification systems processing 100K+ face pairs daily at Kotak Bank, to retail staff-optimization models, to medical imaging pipelines that saved 30 hours/week of diagnostic effort. My work spans the full ML lifecycle: from prototyping pose-estimation models in research labs to shipping OCR systems that cut manual review by 30% in production KYC workflows.Recent Impact:1. Computer Vision at Scale: Optimized face-verification for Kotak Bank\'s KYC pipeline, improving ROC-AUC from 0.91→0.96 with sub-second latency on 100K+ face pairs2. Demand Forecasting: Built LightGBM models across 20K+ SKUs that reduced stockout-driven revenue loss by 14%3. Enterprise AI: Deployed RAG system with dense retrieval + cross-encoder reranking, cutting legal review workload by 42%4. Inventory Optimization: Developed graph neural networks for multi-store allocation, reducing stockouts by 12% for Nestlé across 50+ outlets5. Research: Currently evaluating pose-estimation models (LDPose, AlphaPose, SAM3D) for assistive motion analysis at UW. Technical Edge:I don\'t just train models, I ship them. My work includes model evaluation frameworks, threshold calibration, error analysis, A/B testing, and monitoring pipelines. I\'ve worked with YOLOv7, TrOCR, CNNs, LSTMs, GNNs, and Transformers in production environments. Current Focus: Federated learning systems, Vision Models, RAG architectures for domain-specific retrieval, and multimodal AI applications.Background: Before grad school, I published research on Devanagari character recognition and built everything from autonomous robotics systems to CNN-powered waste segregation deployed with municipal workers - because the best ML solves real problems for real people.I\'m seeking Summer 2026 internships in Machine Learning, Computer Vision, or Applied AI (LLMs/RAG)- where I can bring production-ready ML thinking from day one.View my work on my Portfolio/GitHub attached below.

Experience

1. Graduate Research Assistant

   Paul G. Allen School Of Computer Science & Engineering

   Nov 2025 — Present

Education

• KJ Somaiya College of Engineering, Vidyavihar

  Bachelor of Technology, Electronics and Telecommunication Engineering

• University of Washington

  Master of Science - MS, Data Science