Research Analyst/ Engineer: AI/ML

Institution: National Healthcare Group Family Group: Administration Research Analyst/ Engineer: AI/ML Job Description: Hybrid Architecture Development: Responsible for designing the communication architecture between the client (Windows) and the cloud or local backend, managing complex business logic and system workflows. AI Model Integration: Deploy and optimize Automated Speech Recognition (ASR) models to accurately recognize various accents (e.g., Singaporean) , and utilize Computer Vision to monitor patient posture and ensure the proper use of the eye occluder. Multi-display Interaction Design: Orchestrate real-time, event-driven state machines to ensure seamless state synchronization across multiple displays, specifically the patient-facing touchscreen and the Snellen chart monitor placed at a 4-meter distance. Hardware Coordination: Integrate and maintain data links and drivers for various peripherals, including NRIC scanners, thermal printers, high-resolution webcams, and microphones. Data Security & Compliance: Ensure secure storage of patient test results and performance logs in the cloud (Amazon S3) , implementing strict data anonymization to prevent the retention of sensitive personal information. Job Requirements: 1. Education & Background Studied/Studying in Computer Science, Software Engineering, Computer Engineering, or a related technical field. Strong interest in learning analytical and problem-solving skills Ability to work in a fast-paced environment and manage multiple priorities 2. Front-end Development JavaScript Development & Engineering: Solid proficiency in JavaScript (ES6+) with the ability to manage application logic independently. Familiar with modular development and modern build tools (e.g., Webpack, Vite). Real-time Media APIs: Extensive experience with browser-based media handling, specifically MediaDevices for camera/microphone access and WebRTC/WebSocket for low-latency streaming. Client-side AI & WASM: Familiarity with ONNX Runtime Web to implement browser-side Voice Activity Detection (VAD). Hybrid Rendering: Experience with server-side templating engines (e.g., Jinja2, EJS, or similar). Ability to bridge the gap between server-rendered content and dynamic client-side updates via Vanilla JS and DOM manipulation. Hardware Integration: Experience integrating third-party SDKs (e.g., thermal printer SDKs) into web-based workflows. 3. Back-end & AI Engineering Python Engineering: Expert in modern Python development. Skilled in building robust backend services and managing application logic with a focus on code quality and efficiency. Audio & ASR Inference: Practical experience in deploying and optimizing mainstream ASR frameworks (e.g., Whisper, FunASR, or similar). Familiar with offline ASR workflows and capable of tuning inference performance for specific use cases. Computer Vision (CV): Hands-on experience with MediaPipe for facial landmark, iris detection, and contour extraction. Proficiency in OpenCV for specialized tasks like ArUco marker detection and adaptive image preprocessing. State Management & Synchronization: Expertise in designing event-driven state machines. Proven ability to ensure synchronized state across multiple displays (e.g., Touchscreen vs. VA Display) using tools such as Redis, WebSockets, or other real-time communication protocols. 4. AWS for Data Storage: Practical experience with Amazon S3 for secure data persistence. Proficiency in Boto3 for backend data uploading and AWS SDK (v3) for frontend data retrieval. Familiar with IAM roles and policies to ensure secure access to patient data. 5. AWS Skills for Cloud Computing Amazon Kinesis Video Streams (WebRTC): Hands-on experience building and troubleshooting KVS WebRTC pipelines, including Signaling Channels, role separation (MASTER/VIEWER), and signaling workflows (SDP offer/answer, ICE candidates). Cloud Integration: Skilled in AWS SDKs (Boto3, JS v3) and IAM security best practices (Credential management, secure signaling). Operational Awareness: Comfortable managing region/config via environment variables and deployment settings (e.g., AWS_REGION, credentials, proxy config), and interpreting AWS-side failures from logs/metrics for rapid root-cause analysis.