Satvik Daga
Lead Data Scientist
I architect production-grade AI/ML systems at scale — from semantic search and recommendation engines to MLOps infrastructure on Kubernetes.
NOW
- Building: hybrid semantic search for e-commerce
- Scaling: cross-platform product monitoring with computer vision
- Exploring: demographic prediction for cold-start personalization
Updated 2026-04-22
RECENT WRITING
all posts →PROJECTS
Name-Based Demographic Predictor
CLI tool that predicts customer gender and religion from name with 92% accuracy, integrated into recommendation and marketing pipelines for cold-start personalization.
Hybrid Lexical-Semantic Search
Integrated semantic retrieval into Snapdeal's lexical search pipeline, lifting MRR from 0.39 to 0.63 and NDCG from 0.48 to 0.72.
Intelligent Search System
Semantic search powered by Hugging Face transformers and LangChain with Elasticsearch as a vector database, handling 100K+ daily queries with 40% fewer irrelevant results.
ML Infrastructure Pipeline
End-to-end MLOps platform on AWS EKS with Terraform infrastructure-as-code and Argo CD GitOps, achieving 99.9% uptime and 50% faster deployments.
Cross-Platform Product Monitoring
Image-recognition automation framework that matches Snapdeal products across competing e-commerce platforms, achieving 85% match coverage and surfacing pricing gaps at scale.
Advanced Recommendation Engine
Dual recommender systems combining collaborative filtering and graph embeddings, serving 2M+ monthly active users with 20% engagement lift.
EXPERIENCE
CONTACT