Senior AI/ML Developer

Architected and deployed a production LLM-powered chatbot backend using Java Spring Boot and Python integration layer, serving millions of daily user interactions with sub-500ms latency. Built an end-to-end RAG pipeline for intelligent search and Q&A using LangChain, Hugging Face embeddings, and vector databases. Designed a high-throughput real-time fraud detection platform leveraging Kafka streams, Apache Spark, and PyTorch models. Architected MLOps infrastructure using Kubeflow, MLflow, Airflow, Docker, and Kubernetes. Optimized inference performance using TensorRT, ONNX conversion, quantization, and pruning. Built distributed data engineering pipelines with Apache Spark and Kafka. Delivered an LLMOps platform with standardized deployment templates and automated evaluation dashboards. Engineered microservices using Spring Boot and FastAPI with event-driven architecture and Redis caching. Implemented observability stack with Prometheus, Grafana, Elasticsearch, and drift detection. Led cross-functional collaboration and mentored junior engineers.

Developed LLM-powered chatbot backend services using LangChain, Hugging Face models, and vector search integration. Built a hybrid recommendation engine combining PyTorch embeddings, ML ranking models, and content-based filtering. Designed real-time data pipelines using Kafka and Apache Spark. Implemented LSTM-based anomaly detection models and deployed them with MLflow, Docker, and Kubernetes. Fine-tuned domain-specific LLMs using LoRA/QLoRA and optimized inference latency through ONNX and TensorRT. Deployed ML services to AWS SageMaker and GCP Vertex AI and implemented drift detection dashboards. Collaborated with backend teams to integrate ML inference outputs into microservices.

Developed and optimized machine learning models using Python, scikit-learn, PyTorch, and TensorFlow. Built data preprocessing and feature engineering pipelines with Pandas, NumPy, and Spark. Created inference services using FastAPI and Docker. Performed model evaluation, hyperparameter tuning, and performance tracking using MLflow. Collaborated with data engineers, backend developers, and product teams to design end-to-end ML solutions.