Architected distributed model training and inference system.

Built model serving backend with task scheduling, handling and monitoring system.

Engineered large-scale automated model onboarding pipeline with deploy and inference workers.

Developed finetuning and serving docker for PyTorch, Tensorflow and vLLM models.

Executed data labeling, benchmarking and evaluation for key custom models.

Led critical security initiatives to ensure user data security across products.

Languages: Python, Java, C++, TypeScript, SQL

AI/ML: PyTorch, Tensorflow, vLLM, MLops, LLM serving, Feature Engineering

Infrastructure: Backend APIs, Cloud Computing (GCP), Kubernetes, Docker Management

Tools: Git, Jupyter

Developed Natural Language Understanding model-training pipelines for Alexa/Echo.

Implemented training interface and workflow on remote GPU clusters for distributed computing.