Edge AI Intern, Summer 2026

Research and implement techniques to compress state-of-the-art models (LLMs, CNNs) for edge deployment, utilizing methods such as quantization (INT8, INT4), pruning, and knowledge distillation.,Prototype and deploy inference engines on local hardware (e.g., mobile CPUs/NPUs, edge servers, or embedded systems) using frameworks like ONNX Runtime, TensorFlow Lite, or ExecuTorch.,Analyze and profile model performance to identify bottlenecks; optimize inference pipelines for real-time financial applications (e.g., fraud detection, biometric authentication).,Experiment with split computing strategies to intelligently divide workload between the edge device and the cloud.,Develop rigorous testing suites to measure power consumption, memory footprint, and inference speed across different hardware targets.

Currently enrolled in a Master's program or advanced Undergraduate in Computer Science, Electrical Engineering, or a related field.,Strong proficiency in Python and C++ (specifically for high-performance inference).,Deep understanding of deep learning frameworks (PyTorch or TensorFlow) and their internal mechanics.,Experience with model compression techniques (Quantization, LoRA, etc.).,Familiarity with edge inference runtimes (e.g., ONNX, TensorRT, CoreML, or TFLite).,Research experience or publications in efficient deep learning or systems for ML (Nice-to-have).,Experience with LLM inference optimization (e.g., vLLM, llama.cpp) (Nice-to-have).,Knowledge of hardware-software co-design (understanding how memory hierarchy affects AI performance) (Nice-to-have).,Previous exposure to the financial industry or privacy-preserving technologies (e.g., Federated Learning) (Nice-to-have).

Master's program or advanced Undergraduate