Cloud Infra - Deep Learning Platform

• Distributed LLM Inference (AiBrix)

• Distributed Caching System (RDMA, Fuse, Splice)

• Linux AMD High Performance Computing

• CUDA / NCCL

• Kubernetes

• Cloud Native / AI Native

• Model Context Protocol, AI Agent

Game Development

• Architected robust training pipelines for a GPT-based customer service application

• Integrated Langchain Embeddings and Faiss with OpenAI APIs in response generation

• Engineered a Full-Stack application with a containerized Flask backend on AWS Lambda

• Managed end-to-end automation of software delivery process with CI/CD principles using AWS CodePipeline, reducing deployment time by 30%

• Orchestrated containerized microservices using Kubernetes on AWS EKS, optimizing resource allocation and ensuring seamless autoscaling

• Conceptualized and implemented front-end components with React.js

• Successfully amplifying user engagement and augmenting dwell time by 25%

• Boosted data collection and retrieval efficiency by 25% through database schema optimization

• Led development of user authentication and authorization with Google Firebase

• Deployed a decision-tree recommender system on AWS Lambda, achieving a stateless design that scaled to handle 10,000+ requests per minuteonducted in-depth analysis of complex data sets to identify trends and patterns

• Applied AWS Elastic Load Balancing to the decision-tree system, enhancing traffic distribution and reducing response times by 25%

• Leveraged AWS S3 and DynamoDB for state management, ensuring 99.9% data consistency and resilience across all instances

• Established automatic replication processes, resulting in a 99.95% uptime and swift failover during peak usage times