Harsh Kumar

• Built the core backend infrastructure from scratch for the social application using Go, MongoDB, and GCP, handling user authentication, connection logic, and geo-location features.
• Engineered a separate real-time chat microservice using WebSockets and Redis Pub/Sub, implementing handlers for messages, images, reactions, and read receipts, integrated with Firebase for push notifications.
• Developed critical auxiliary services, including a secure OTP verification system using Zixflow and Redis for session management, and implemented a multi-key rate limiter to protect API endpoints.

• Engineered the complete backend for FindCreators.ai, delivering 45+ RESTful APIs using Go (Fiber) and MongoDB on GCP.
• Developed a high-efficiency Selenium web scraper to aggregate over 30,000 influencer emails within 30 days, directly fueling the successful onboarding of 900+ creators to the platform.

• Engineered a financial data reconciliation pipeline on Databricks using PySpark, implementing over 20 validation logics for asset management clients that eliminated manual processing and ensured data integrity across systems.
• Designed a modular architecture for data ingestion, processing, and reporting, which included calculating fund IRR with SciPy and Pandas and persisting reconciled data to Spark tables to facilitate BI reporting and analysis.
• Architected a reusable data quality framework in Python, utilizing a dynamic YAML configuration to validate complex Excel schemas and custom business rules, creating a robust gatekeeper to prevent data corruption in downstream pipelines.

• Built a real-time chat support feature using Apollo GraphQL, React.js, Redux Toolkit, and Tailwind CSS, enhancing user communication.
• Improved admin dashboard UI responsiveness and fixed bugs using Tailwind CSS.

Freelance

• Engineered the complete backend infrastructure for SalesOutreach.io, integrating Stripe for payments and premier data providers like Wiza and implemented email outreach using Email Bison.
• Implemented a credit and atomic transaction system using Go and MongoDB transactions to ensure data integrity and reliability for automated sales outreach.

• Achieved v1.26 feature parity by integrating support for customMetadata in document operations, ensuring alignment with the core engine capabilities.
• Engineered comprehensive unit and integration tests to validate metadata transmission, ensuring robust data handling and system stability.