Jay
Prajapati

Built a personalized inbox and notification platform inside the MyMicrochip dashboard that delivers user-specific updates, secure document activity, campaign communication, and platform events through a centralized experience. A large part of the work involved designing backend workflows that merged data across multiple systems, handled document snapshot versioning, campaign eligibility, deduplication, and async processing while keeping the admin tooling flexible enough for business teams to manage templates, workflows, and notification behavior without engineering dependency for every operational change.

Worked on an offline-first product discovery platform handling 60K+ products with focus on fast client-side search, aggressive caching, and large dataset handling across complex filtering workflows. Built offline support for the PWA experience using IndexedDB-based browser caching and background sync, allowing the application to behave more like a desktop tool with continued usability even under unreliable or completely disconnected network conditions.

Working on an AI assistant focused on document retrieval, product discovery, and dashboard workflows across Microchip platforms. The project involves building agentic flows using MCP integrations to let users perform actions through natural language interactions, including document search, support workflows, notification subscriptions, and personalized dashboard management using retrieval pipelines, vector search, and contextual orchestration.

Working on an NLP-driven product search and recommendation platform for structured engineering datasets, allowing users to query products naturally for competitor alternatives, similar Microchip parts, category filtering, and hybrid parametric search. The workflow combines intent classification, regex preprocessing, RAG retrieval, LLM-guided SQL generation, reranking, and scoring to deliver contextual search results through conversational interactions.