AI Model Training and Optimization Engineer

AI Model Training and Optimization Engineer

We are seeking a skilled AI Model Training & Optimization Engineer to join our AI / ML team. In this role, you will be responsible for fine-tuning machine learning models, optimizing hyperparameters, and scaling inference to ensure high performance and efficiency across production environments. You will work closely with data scientists, MLOps engineers, and infrastructure teams to deliver state-of-the-art solutions at scale. Key Responsibilities :

Fine-tune pre-trained models (NLP, CV, or other domains) to suit specific business needs and datasets. Optimize hyperparameters to improve model accuracy, speed, and generalization. Perform inference optimization using tools like TensorRT, ONNX Runtime, and other acceleration toolkits. Collaborate with engineering teams to deploy models across GPU-accelerated environments. Analyze model performance across various hardware configurations and recommend improvements. Maintain scalable, reusable training pipelines using frameworks like PyTorch Lightning or TensorFlow Extended (TFX). Convert and optimize models for cross-platform inference using ONNX, TorchScript, or TensorFlow Lite. Stay updated on advancements in model compression, quantization, and low-latency inference techniques. Required Qualifications :

Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, or related field. 3–5 years of hands-on experience in training, fine-tuning, and optimizing machine learning or deep learning models. Strong proficiency in TensorFlow and PyTorch. Experience with model export formats such as ONNX, TorchScript, or TF Lite. Familiarity with inference optimization libraries like TensorRT, ONNX Runtime, or OpenVINO. Solid understanding of GPU architecture and performance profiling tools (e.g., NVIDIA Nsight, nvprof). Experience working in cloud-based or containerized environments (AWS / GCP / Azure, Docker, Kubernetes). Proficient in Python and common ML / DL libraries (NumPy, Pandas, Scikit-learn). Preferred Qualifications :

Experience with model quantization, pruning, or distillation. Understanding of MLOps workflows and deployment practices. Exposure to distributed training frameworks (e.g., Horovod, DDP). Experience in benchmarking and profiling inference performance in real-time systems.

#J-18808-Ljbffr