• Port and deploy models across platforms (CPU, GPU, edge devices using ONNX, TensorRT, TFLite, etc.)
• Benchmark models and analyze trade-o s between accuracy, speed, and resource usage
• Debug model behavior and improve robustness in real-world scenarios to integrate models into production systems
• Maintain clear documentation of experiments, model versions, and performance metrics
• Stay updated with advancements in computer vision and applied ML
• Design, train, and evaluate deep learning models for computer vision tasks (classification, detection, segmentation)

• Work with advance CNN architectures and optimize models for performance, latency, and memory (quantization, pruning, fp16, batching strategies) • Handle end-to-end pipelines including data preprocessing, augmentation, training, validation, and testing

• Port and deploy models across platforms (CPU, GPU, edge devices using ONNX, TensorRT, TFLite, etc.)
• Benchmark models and analyze trade-o s between accuracy, speed, and resource usage

• Debug model behavior and improve robustness in real-world scenarios to integrate models into production systems • Maintain clear documentation of experiments, model versions, and performance metrics

• Stay updated with advancements in computer vision and applied ML

Requirements

Bachelor’s or Master’s degree in Computer Science, AI, Electrical Engineering, or related field
4+ years of hands-on experience in AI/ML, with strong focus on computer vision
Solid understanding of CNNs and deep learning fundamentals
Strong experience with frameworks like PyTorch or TensorFlow
Good programming skills in Python
Experience with model training, evaluation, and hyperparameter tuning
Practical experience in model optimization techniques (quantization, pruning, distillation)
Knowledge of edge AI constraints and deployment challenges
Experience in model deployment and porting (ONNX, TensorRT, OpenVINO, TFLite, etc.)
Familiarity with Linux environments and basic system-level understanding
Experience with version control systems like Git

Good to Have

Exposure to NLP concepts or transformer-based models
Experience working with multimodal models (vision + text)
Familiarity with CUDA or hardware acceleration techniques
Experience with MLOps practices (experiment tracking, CI/CD for models)

What We’re Looking For

Strong problem-solving mindset with attention to real-world constraints
Ability to go beyond model accuracy and think in terms of system performance
Clear communication and ability to work across teams