• Designed and implemented a natural language–to–SQL retrieval pipeline using LLMs and RAG, enabling internal teams to query orders, drivers, and payment data conversationally instead of writing complex SQL, reducing operational support overhead and improving data accessibility

• Built scalable TypeScript REST APIs to ensure reliable order processing, real-time notifications, and safe refund handling for a high-volume delivery platform serving 100K+ daily users and 1M+ orders, preventing inconsistent states and improving customer and driver experience

• Designed a Redis-backed dispatch system to match orders with the nearest available drivers and integrated PayPal and Stripe for secure payment processing, reducing delivery delays and ensuring seamless, duplicate-free transactions

• Eliminated SQS/DLQ processing bottlenecks by introducing concurrent Go worker pools and horizontal scaling, increasing throughput to 5K+ orders/sec and significantly reducing backlog-related delivery delays

• Automated CI/CD pipelines using GitHub Actions to eliminate manual deployment steps, reduce release time by 50%, and ensure consistent, error-free deployments across environments

• Implemented JUnit test suites to achieve 80%+ code coverage, reducing production defects and increasing confidence in releases as the platform scaled

• Built and deployed a scalable Generative AI chatbot to address fragmented, hard-to-navigate public water education resources amid Arizona’s escalating drought crisis, transforming static information into an interactive, real-time conversational experience

• Developed using Python (Flask → FastAPI) and integrated GPT-3.5 Turbo and Amazon Bedrock Claude, deployed on AWS Fargate (ECS/ECR) with infrastructure provisioned via Terraform and AWS CDK, delivering a secure, cloud-native production platform that achieved 95% customer satisfaction

• Integrated a Retrieval-Augmented Generation (RAG) pipeline using LangChain and ChromaDB with semantic vector search to improve answer relevance, significantly enhancing response accuracy and user experience

• Implemented multi-layer LLM guardrails—including prompt injection detection flags and user query verification—to address security risks and potential misuse in a public-facing AI platform, ensuring protection against malicious attacks and preserving response integrity

• Integrated real-time audio-to-text translation using Amazon Transcribe and FastAPI WebSockets to solve accessibility and engagement gaps, enabling voice-based interaction and enhancing overall user experience

• Implemented multilingual language detection and translation workflows, expanding accessibility to non-English users and increasing platform adoption

• Integrated the chatbot into a React-based frontend via streaming APIs, resulting in a smoother and more responsive user experience

• Optimized session log storage by transitioning from RDS to a hybrid AWS DynamoDB and S3 architecture to handle real-time data interactions and long-term storage, cutting costs by over 60%

• Designed and deployed a secure network architecture with AWS CloudFront, ALB, and Fargate to enhance scalability and security, leading to efficient traffic management with secure headers and dynamic scaling leveraging Fargate's serverless compute

Client: T-Mobile , ATT, Verizon, Vodafone

• Designed and delivered scalable e-commerce backend platforms through microservices modernization, security enhancements, and performance optimization, significantly improving system scalability, reliability, and search experience

• Led development of a Java, Spring Boot, and Docker–based pricing microservice that consolidated scattered pricing logic into a single service, simplifying a complex pricing system, improving ownership, and enabling faster pricing updates while supporting high traffic during sales and promotions with ~50% better scalability

• Strengthened platform security by implementing an ABAC-based authorization model with elevated role controls across 15 microservices, preventing unauthorized customer data access while enabling secure internal and support operations, reducing security risks by 25%

• Reduced infrastructure cost and improved REST API performance by optimizing JSON serialization in Couchbase using Jackson annotations, decreasing document size by 26% and lowering CPU usage during read/write operations by 10% without requiring production data migration

• Improved system reliability and data consistency by eliminating 30% of redundant message processing through the implementation of an Apache Kafka idempotency mechanism across 90 microservices

• Enhanced the Discovery microservice with Elasticsearch-based faceted search, allowing users to dynamically filter results. This significantly improved relevance and usability, leading to roughly a 40% increase in search engagement metrics.

• Mentored and led a team of 3 interns to perform root cause analysis (RCA) on over 50 e-commerce microservices to identify and rectify defects, yielding the resolution of 90% of critical issues and a significant enhancement in system reliability

• Engineered and maintained Java microservices leveraging Spring Boot, achieving scalable deployments on cloud platforms via Docker and Kubernetes.

• Conducted Knowledge Transition Sessions for new joiners, accelerating their integration and upskilling, which improved team productivity and expedited project timelines.

• Acted as the primary support contact for major US telecom clients, swiftly resolving escalated production issues, which bolstered client satisfaction and reinforced trust in our service reliability.

• Acquired proficiency in Agile methodologies and technologies such as Apache Kafka, Couchbase, ELK Stack, Jenkins, and Maven, applying these practices to enhance project efficiency and technological adaptability.

• Designed and launched new REST APIs based on gathered requirements, providing tailored business solutions that expanded service capabilities and enriched user experiences.