Deep Verma

Architected end-to-end prior authorization platform Integrated LLM-powered document enhancement (Ethermed) with insurance clearinghouse APIs (Waystar), improving claim submission success rate by 45% across 33+ insurance carriers while maintaining 99.9% transactional integrity. Designed and implemented scalable middleware architecture that standardized data ingestion across fragmented healthcare protocols (HL7, FHIR, custom APIs), reducing hospital onboarding time by 40% and enabling seamless integration with disparate EHR systems. Built fault-tolerant bidirectional sync pipelines with idempotent retry logic and asynchronous recovery mechanisms to handle distributed state failures across clinical systems, ensuring zero data loss in production. Led technical implementation using FastAPI, SQLAlchemy, and async patterns with comprehensive test coverage, establishing architectural patterns and best practices adopted across the engineering team.

Founding Engineer - Fashinza (B2B Manufacturing Marketplace Platform). Engineered Finance microservice with domain-driven design principles; implemented event-driven architecture using pub/sub patterns for real-time AR/AP reconciliation, designed RESTful APIs with idempotent endpoints and optimistic locking for concurrent transactions, built streaming data pipelines for financial aggregation using Celery workers, and integrated APM (Application Performance Monitoring) for distributed tracing and performance monitoring—reducing reconciliation latency from O(days) to O(seconds). Designed fault-tolerant document extraction system integrating Amazon Textract with OpenAI GPT-4 API; implemented asynchronous processing with Celery workers, structured prompt engineering with JSON schema validation using serializers, comprehensive error handling with circuit breaker patterns, and retry mechanisms with exponential backoff—reducing processing time of POs from hours to minutes.

Time & Action Plan (TNA): Developed a forecasting engine to predict delivery turnaround time using rule-based and historical data and adjusting timelines for parallel and sequential tasks, thus identifying potential delivery delays. Qapp: Developed a quality control system that digitized manufacturing inspections, defect tracking, and delivered real-time efficiency metrics and DHU calculations across multiple production facilities, thus reducing up to 65% garment rejection rates. Integrated third-party platforms including PayPal, Asana, and Google Sheets for enhanced functionality. Developed unit and integration tests for 50+ APIs using Pytest, ensuring product reliability. Continuously optimized APIs to minimize system latency, maintaining response times below 150ms.

Integrated terminal on to a web app. Developed admin panel alongwith Email, notification, and alarms service using EC2, SNS & SES service to monitor client and system health status.

Development of new customer support web app & displaying huge amount of data on app. Thus increasing the sales team requests handling efficiency per hour by 4x.