Nitish Joshi

• Optimized internal graph algorithm which led training data finding time from 20 mins to 1 seconds on 5 million records
• Designed and implemented internal licensing system for enterprise code, ensuring controlled access to features
• Refactored Java code base to adhere to Object-Oriented Programming (OOP) principles and SOLID design principles, enhancing maintainability and scalability
• Implemented feature to incorporate scoring for relationships among entities in household domain data clustering
• Designed and implemented feature to explain match results by generating various network graphs using Python d3graph library, improving data visualization and insights.

Distributed Authentication Architecture

Designed central authentication for Honeywell public requests to ensure security, scalability and high availability.

Monolith Spark Application To Java Migration

Migrated Databricks spark job to Java-based microservice architecture. During this transition, streamlined standard APIs were extracted from monolithic system, resulting in substantial cost savings.

Centralized Version Manager

Designed and developed centralized Version Manager enhanced system performance by implementing efficient caching through Redis, resulting in reduction in network call latency.

Search API

Developed a search API that efficiently retrieves documents from Elasticsearch and seamlessly delivers them to frontend services. This API is equipped with features such as aggregation, fuzzy search, handling of time-series data, and more, ensuring robust and versatile document retrieval capabilities. Introduced in-memory caching to optimize API response times, resulting in a noteworthy reduction of 10 milliseconds.

Onboarding

Introduced multithreaded batch processing feature to expedite asynchronous customer onboarding process, improving efficiency and speed.

• Suggested and executed integration of Redis for pagination, enhancing system performance and user experience.
• Revamped integration tests, resulting in 15-minute reduction in build time.