Ashwin Kumar J S Email & Phone Number

Bengaluru, Karnataka, India

Expanded LLM services to multiple languages by implementing a real-time translation layer,enabling seamless integration and adaptation for a wide range of linguistic variations andregional nuances.Spearheaded extensive experimentation and research initiatives, employing cutting-edge frame-works and optimization techniques and successfully reduced the operational costs of large language models while maintaining performance and scalability.

Bengaluru, Karnataka, India

Productionizing Machine Learning Models: Played a pivotal role in the deployment and productionization of machine learning models, ensuring seamless integration into the company's systems. This involved containerization using Docker for efficient scalability and management.Developing REST APIs: Designed and developed RESTful APIs to facilitate data exchange and communication between various services and clients, enhancing system interoperability and user experience.Kafka Topic Integration: Successfully integrated Kafka topics into the data architecture to enable real-time event-driven communication and data streaming, improving data processing efficiency.Websockets Implementation: Implemented Websockets for real-time communication between clients and the server, enhancing user engagement and interactivity within the application.GitHub CI/CD Pipeline: Established a robust Continuous Integration/Continuous Deployment (CI/CD) pipeline on GitHub, automating the build, testing, and deployment processes, leading to faster and more reliable software releases.Optimizing Large language models using SOTA frameworks for production environmentsCollaborated cross-functionally with engineering and data science teams to fine-tune models for maximum accuracy and performance.

Deep Learning Model Benchmarking:Conducted extensive benchmarking of deep learning models using TensorFlow, PyTorch, and OpenVINO on leading cloud vendor machines. Analyzed and compared model performance, inference speed, and resource utilization to optimize model selection and deployment strategies.Airflow Pipeline Jobs in Google Composer:Managed and optimized data workflows by designing and orchestrating complex data processing pipelines using Apache Airflow within Google Composer. Ensured data reliability, scalability, and scheduling for critical business operations.CI/CD-enabled MLOps Pipeline in AWS SageMaker:Led the design, development, and deployment of a robust CI/CD-enabled MLOps pipeline in AWS SageMaker. Streamlined the model training, testing, and deployment process, resulting in faster model updates and improved model lifecycle management.CI/CD-enabled MLOps Pipeline in Kubernetes Cluster with Load Balancing:Architected and implemented a CI/CD-enabled MLOps pipeline in a Kubernetes cluster with load balancing. This enabled automated scaling, efficient resource allocation, and high availability of machine learning applications, ensuring optimal performance and resilience.

Student Body

Technical Events