• Designed and implemented an AI-powered chatbot for university students, leveraging large language models (LLMs) to answer academic, administrative, and campus-related queries.

• Integrated the chatbot with university portals and databases, enabling real-time access to course schedules, exam results, and campus events.

• Utilized NLP techniques and prompt engineering to improve response relevance, achieving over 90% accuracy in addressing student inquiries.

• Collected user feedback and applied continuous learning pipelines to fine-tune the chatbot, enhancing user satisfaction and engagement.

I played a pivotal role in developing scalable API infrastructure at Accenture, ensuring high availability and performance during peak traffic.

• Engineered a high-performance RESTful API using Java and Spring Boot to serve real-time transaction data, providing personalized spending insights for millions of users.

• Integrated a rule-based categorization engine to automatically map raw merchant data into financial categories with a 99.9% classification accuracy rate.

• Optimized system performance by utilizing Redis for frequent data lookup caching, reducing backend database load by 35%.

• Enhanced user experience by improving UI data retrieval speeds by 30% through strategic caching and API optimization.

• Leveraged Kafka to ingest and process 1.2M+ daily transactions, ensuring synchronized account analytics with sub-second latency.

• Deployed and managed scalable API infrastructure on AWS using Docker and Terraform to ensure high availability.

• Architected the system to handle 3x surges in traffic during peak transaction periods, maintaining stability and performance.

I played a pivotal role in enhancing a local networking platform's performance and reliability.

• Developed Java Spring Boot REST APIs, supporting real-time chat and user interactions for 10K+ users.

• Utilized Docker for microservices, improving deployment efficiency and scalability during peak usage.

• Integrated Apache Kafka for processing over 5K events per minute, ensuring low-latency delivery.