Edge Ai Engineer

Position: Edge AI Engineer - Cancard Inc Location: New Delhi / Toronto Role Type: Full Time Cancard Inc and Advaa Health are seeking an experienced, engaged, and hands-on healthcare industry Edge Computing / IOT / CV expert for the role of Edge AI Engineer. This role will be pivotal in driving development, launch, and successfully commercializing an innovative portfolio of digital healthcare products for global markets. Cancard Inc has been a multi-technology company based in Markham (Toronto) since 1989. Both Cancard and its sister company, Advaa Health, are at the forefront of transforming primary healthcare through technological innovation. Our mission is to empower primary care physicians with state-of-the-art digital tools and solutions that streamline their practices and significantly reduce administrative burdens and operational costs. Amidst increasing paperwork and complex administrative tasks faced by healthcare professionals, we serve as a key partner, enabling physicians to focus on patient care. Our healthcare product portfolio leverages cutting-edge technologies in data analytics, artificial intelligence, and cloud computing to offer seamless, intuitive, and cost-effective solutions. By integrating our systems, primary care practices can enhance patient engagement, optimize appointment scheduling, automate billing and coding processes, and access comprehensive patient health records in real-time. These advancements not only improve the quality of care provided but also contribute to a significant reduction in overhead costs. This position offers a unique opportunity for AI Engineers who are passionate about solving critical healthcare challenges to learn and grow within the company. The role provides direct experience and exposure to customers in the US, Canada, and other global markets. KEY RESPONSIBILITIES: EDGE COMPUTING MODEL DEPLOYMENT: - Convert, optimize, and deploy trained AI/ML models for edge inference using tools such as: o TensorRT (for NVIDIA Jetson and GPUs) o ONNX Runtime (for hardware-agnostic and cross-platform deployment) o OpenVINO (for Intel-based edge devices) o TFLite (for mobile and microcontroller-based healthcare applications) - Apply model optimization techniques including: o Post-training quantization (INT8, FP16) o Pruning and weight clustering o Model conversion to ONNX or IR formats for interoperability - Benchmark models for: Latency, Throughput, Memory Footprint, and Power Efficiency - Profile and debug model performance using tools such as: NVIDIA Nsight, TensorRT Profiler, Intel DevCloud, Netron, or Edge TPU Compiler - Integrate AI inference engines with real-time edge pipelines, ensuring low-latency decision-making in clinical or diagnostic environments. IOT SENSOR INTEGRATION: - Integrate and interface with a wide range of biomedical and motion sensors commonly used in healthcare devices: o Inertial Measurement Units (IMUs): Accelerometers, gyroscopes, magnetometers. o Physiological Sensors: Heart rate (PPG), SpO2, ECG, EMG, temperature, respiration, and blood pressure sensors. o Environmental Sensors: CO₂, humidity, light exposure (for ambient health monitoring). - Sensor Communication Protocols: o Work with I²C, SPI, UART, BLE (Bluetooth Low Energy), and USB HID for data acquisition and transmission. o Implement sensor drivers, data parsing, and calibration routines for real-time use. - Data Fusion & Preprocessing: o Apply filtering (Kalman, low-pass, high-pass) and sensor fusion algorithms to combine multiple sensor modalities for robust signal interpretation. o Normalize and preprocess raw sensor data for efficient input into ML models. - Time-Series Analysis: o Manage high-frequency data streams and maintain synchronization across multiple sensors. o Build and deploy real-time AI pipelines for activity recognition, fall detection, gait analysis, and anomaly detection in patient vitals. - Edge Implementation: o Run lightweight signal processing and classification models directly on microcontrollers or edge devices. o Use platforms such as TinyML, Edge Impulse, or CMSIS-NN for embedded ML on sensor data. - Device Interoperability & Testing: o Ensure compatibility with commercial wearable platforms (e.G., smart bands, chest straps, ECG patches). o Conduct in-field testing and validation in collaboration with clinical or biomedical teams. COMPUTER VISION EXPERTISE: - Strong experience developing and deploying computer vision models for edge devices (e.G., object detection, segmentation, classification, pose estimation). - Proficient with deep learning architectures used in computer vision, such as CNNs, MobileNet, EfficientNet, YOLO, U-Net, and Vision Transformers (ViT). - Hands-on experience with optimizing CV models for performance on resource-constrained devices (e.G., quantization, pruning, knowledge distillation). - Experience using OpenCV and other image processing libraries in Python and/or C++ - Familiarity with real-time video analytics, multi-frame inference, and hardware-accelerated vision pipelines. - Understanding of image preprocessing techniques for low-quality or noisy medical data (e.G., denoising, contrast enhancement, resizing). - Experience integrating camera systems and working with edge hardware with GPU/TPU/ISP acceleration. DEEP LEARNING EXPERTISE (Preferred): - Solid experience designing and implementing deep learning models for edge deployment, with emphasis on applications in hospital environments. o Knowledge of developing and optimizing models using Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM). o Transformers and attention-based architectures (e.G., Temporal Fusion Transformers, Vision Transformers) for capturing long-range dependencies in sequential health data. - Experience deploying deep learning models on edge inference platforms (e.G., TensorFlow Lite, ONNX Runtime, NVIDIA TensorRT, OpenVINO). - Strong understanding of latency, throughput, and memory trade-offs in edge scenarios, especially for real-time clinical decision support. MODEL OPTIMIZATION & DEPLOYMENT: - Design, optimize, and deploy machine learning models for real-time inference on edge devices in clinical and hospital environments. - Implement quantization, pruning, and compression techniques to ensure efficiency on resource-constrained platforms. - Port AI models to run on hardware such as NVIDIA Jetson, Intel Movidius, ARM Cortex-M, or custom medical edge hardware. HEALTHCARE INDUSTRY SOLUTIONS: - Work on edge AI solutions in medical imaging (e.G., X-ray, CT, MRI), patient vitals monitoring, telemedicine devices, and wearable health trackers. - Ensure models comply with healthcare standards, data formats such as DICOM, HL7, and FHIR. - Integrate AI into point-of-care diagnostics, clinical decision support systems, or edge-assisted robotic surgeries. SOFTWARE & HARDWARE INTEGRATION: - Interface with camera modules, biosensors, and physiological data sources (e.G., ECG, SpO2). - Build low-latency pipelines that support real-time analysis in mission-critical use cases. - Collaborate with hardware engineers to ensure compatibility and performance across different boards and chipsets. VALIDATION & COMPLIANCE: - Ensure solutions meet regulatory requirements (e.G., FDA, CE, ISO 13485, IEC 62304). - Collaborate on validation protocols, clinical testing, and explainability in AI decisions for medical accountability. CROSS-FUNCTIONAL COLLABORATION: - Work closely with data scientists, clinical SMEs, software engineers, QA testers, and others. - Document all processes for traceability and regulatory submission. QUALIFICATIONS: Education: - Bachelor’s or Master’s degree in Computer Science, Electronics, or related fields. Technical Skills: - ML Frameworks: PyTorch, TensorFlow, Keras, ONNX. - Edge Tools: TensorRT, TFLite, OpenVINO, TVM. - Programming Languages: Python, C++, C. - Embedded Experience: Linux-based OS (Yocto, Ubuntu), RTOS, and edge platforms (Jetson, NXP i.MX, Intel NUC, Coral). - Medical Data Standards: Exposure to DICOM, HL7, FHIR - Hardware Acceleration: CUDA, OpenCL, ARM NN, DSP/TPU integrations. - Device Communication: Bluetooth LE, USB, UART, SPI, I2C for connecting medical peripherals. - Security & Privacy: Familiarity with HIPAA, data encryption, secure firmware update practices. Soft Skills: - Passion for healthcare innovation. - Ability to translate clinical requirements into technical solutions. - Clear and concise documentation and communication. - Strong ownership and a proactive, detail-oriented mindset. Preferred Qualifications: - Experience developing AI for digital pathology, radiology, or wearable health devices. - Familiarity with clinical workflows and hospital IT systems (PACS/RIS). - Hands-on experience in federated learning or on-device learning for privacy-preserving AI. - Prior involvement in FDA/CE Class II or III device development lifecycle. WHAT WE OFFER: - Competitive salary and benefits package. - Flexible working hours. - A dynamic and supportive work environment with opportunities for professional growth and development. - The chance to work on meaningful projects that have a real impact on healthcare. HOW TO APPLY: Please submit your resume, cover letter, and any relevant work samples or project portfolios to In your cover letter, explain why you're interested in this role and how your background and experience make you a good fit for our team. We thank all applicants for their interest in joining Cancard, but only those selected for an interview will be contacted. Cancard is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.