Looking for a job

Ujjwal Baranwal

@ujjwalbaranwal

Message

Software Engineer at KUBRA building billing-support AI that reduced routine inquiry handling time by 24%.

United States

Message

At KUBRA, I built a Generative AI assistant for billing-support teams that reduced average handling time for routine inquiries by 24%. I also developed retrieval pipelines for utility FAQs and billing policies, improving Recall@5 by 12 percentage points.

I connected AI agents to account and payment services through Java and Spring Boot microservices, automating more than 80 daily lookups. By optimizing context, token budgets, and retrieval caching, I cut p95 LLM response latency by 31% while maintaining evaluation pass rates above 92%.

At Synchronoss Technologies, I improved the reliability and performance of high-volume Java ingestion and REST API systems. My projects also include a domain-specific legal chatbot, a real-time network monitoring system, and a secure peer-to-peer messenger.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer - AI/LLM Integration

Current

KUBRA

Feb 2025 - Present (1 year 8 months)

Built a Generative AI assistant for billing-support teams using Python, FastAPI, Claude, and Amazon Bedrock, reducing average handling time for routine inquiries by 24%. Developed RAG pipelines and connected AI agents to account and payment-status services through Java and Spring Boot microservices.

Python fastAPI Claude Amazon Bedrock RAG Java Spring Boot REST APIs OAuth Redis AWS ECS Docker GitHub Actions Cloudwatch Structured Outputs Prompt Engineering GuardRails LLM Evaluation CIS Connectors JSON schema

Software Engineer

Synchronoss Technologies

May 2023 - Jul 2024 (1 year 2 months)

Reworked a high-volume Java Spring Boot ingestion pipeline processing ~1.5–2M records daily, enforcing idempotency and improving reliability. Optimized REST API resilience and AWS ECS Auto Scaling, reducing P95 latency and improving database performance.

Java Spring Boot PostgreSQL Oracle AWS ECS Docker Jenkins AWS CodePipeline HikariCP REST APIs Idempotency Retry Strategies AutoScaling ALB Cloudwatch SQL Query Optimization Database Indexing Connection Pooling

Full-Stack Developer

OneCommunity Global

Strengthened platform reliability by reviewing pull requests and resolving production incidents. Built reusable React and TypeScript components and optimized MongoDB schemas for complex user-data relationships.

React TypeScript MongoDB NodeJS Unit Testing Integration Testing Code Review Design Systems

Generative AI Developer

Sayaam

Architected and deployed a production-ready serverless Generative AI chatbot using Python, AWS Lambda, and API Gateway, achieving sub-200ms average response latency. Integrated Groq and Gemini APIs with structured prompt engineering and schema-driven MongoDB retrieval.

Python AWS Lambda API Gateway GROQ Gemini Prompt Engineering MongoDB GitHub Actions AWS Systems Manager Parameter Store Serverless CI CD

Student Developer

Paqt

Integrated Google Calendar API for real-time meeting coordination and developed AI-assisted Zapier workflows to automate document preparation. Built an end-to-end document workflow connecting PHP Laravel APIs with Node.js services.

Google Calendar API Zapier PHP Laravel NodeJS API Integration Document Management