Aryan Raj

Software Engineer with 3.5+ years of experience building scalable microservices and REST APIs using Java and Spring Boot, now focused on AI-powered applications using LLMs and agentic workflows. Hands-on experience with LangChain, LangGraph, RAG pipelines, prompt engineering, and LLM integrations (OpenAI, Claude) along with strong expertise in distributed systems, caching, async processing and cloud-native architecture.

Software Engineer at Capgemini Technology Services (2022-10 – Present)

• Worked closely with a cross-functional team to modernize the backend of Ashley — a large-scale home furnishings e-commerce platform — using Spring Boot microservices, improving page load times and reducing API response latency by 40% during peak sale events.
• Integrated Razorpay payment gateway with OTP-based verification and automated notification workflows. Also Implemented JWT authentication with RBAC, ensuring secure and role-based access control.
• Designed and implemented a Redis caching and RabbitMQ queue strategy to manage high-traffic stock updates, reducing order failures by 30%.
• Owned the CI/CD pipeline setup using GitHub Actions and Docker, with AWS EC2/S3 for staging and Lambda for background tasks, enabling zero-downtime deployments.
• Built internal dashboards using React.js to visualize order status, payment events, and stock updates, integrating them with backend APIs for real-time insights.
• Prototyped and integrated LLM-based agentic workflows using LangChain/LangGraph, implementing context-aware processing, multi-step reasoning, and task orchestration using Python and OpenAI/Claude APIs to automate backend workflows and improve system efficiency.

Machine Learning Intern at Indian Institute Of Technology, Kharagpur (2021-01 – 2021-04)

• Collected, annotated, and preprocessed satellite data from two locations (Rosamand and Palmdale) to segment landscapes containing hills, buildings and water bodies.
• Trained Mask R-CNN with ResNet50 and ResNet101 backbones using various augmentation methods, achieving 86% pixel accuracy.
• Utilized KL-Divergence and K-Means to perform urban image segmentation, with a focus on identifying buildings.

B.Tech in Electronics And Communication Engineering – Institute Of Engineering And Management (2018-08 – 2022-07)