Angible is a Taipei-based AI startup founded in 2025, dedicated to accelerating AI adoption in the retail industry through computer vision and edge computing. In our first year, we successfully deployed in-house AI solutions across multiple retail environments spanning Europe, North America, and Southeast Asia, proving our ability to solve real business pain points at scale. The company is now in a critical phase of rapid business expansion and continuous product refinement, focused on building intelligent solutions that improve retail operational efficiency, reduce shrinkage, and enhance the customer experience.

Our team brings deep expertise across AI, edge computing, hardware design, and retail operations, working in a highly collaborative and fast-iterating environment. Driven by the mission to make retail smarter and more efficient, Angible leverages AI and edge computing to help businesses boost profitability and maintain a competitive edge in a rapidly evolving market — making AI a seamless part of everyday operations.

At Angible, we value transparency, clarity of goals, and mutual trust, and we place great importance on each team member's impact and growth. You will have the opportunity to participate directly in building products from 0 to 1 and scaling them beyond, working alongside a team that is both visionary and action-oriented to rapidly validate hypotheses and bring technology to life in real-world scenarios.

We're looking for a Senior ML Engineer who can reliably deploy AI vision models to edge devices in retail stores, achieving scalable product deployment. You will be responsible for end-to-end deployment and operations — from video stream tuning and inference optimization to building MLOps pipelines — ensuring the system runs stably in real-world environments.

You will collaborate with the ML team to ensure smooth training and deployment workflows, align with the backend team on integrating inference results into APIs, and work with SRE / client IT teams to handle networking, device, and remote operations issues at store locations.

• ML: PyTorch, ONNX, OpenVINO, TensorRT
• MLOps: ClearML / MLflow / DVC
• Streaming: FFmpeg, RTSP, WebRTC
• Infrastructure: Docker, GitHub Actions, Linux
• Hardware: NVIDIA GPU, Intel platforms

• Deploy AI to in-store edge devices (NVIDIA / Intel), achieving stable, low-latency, maintainable production systems
• Tune video streaming and inference quality: Optimize RTSP / FFmpeg / WebRTC streaming parameters and OpenVINO / TensorRT inference performance
• Build MLOps and data pipelines: Connect the full pipeline from online data → labeling → training → evaluation → deployment, integrated with CI/CD
• Establish observability and operations mechanisms: Monitor streaming and inference health (packet loss, latency, FPS, temperature, memory), design reporting and auto-recovery mechanisms
• Make technical decisions within business constraints: Find the balance between model accuracy and deployment cost, latency, and stability
• Rapidly identify and fix issues: Across multiple stores, multiple cameras, and various network conditions, quickly locate bottlenecks and resolve them

• Python as primary language, with good software engineering habits (testing, readability, maintainability)
• Deep learning / computer vision hands-on experience: Proficient with PyTorch / ONNX, with experience in object detection / segmentation / MOT / Re-ID; capable of dataset design, evaluation, and model export optimization (FP16/INT8 quantization, ONNX → OpenVINO/TensorRT)
• Video streaming implementation and debugging: Hands-on experience with FFmpeg, RTSP, WebRTC; able to handle latency, packet, and synchronization issues
• Edge inference deployment and tuning: Practical experience with OpenVINO, TensorRT, and NVIDIA platforms for model conversion, deployment, and performance tuning
• Docker and basic CI/CD, along with data pipeline / experiment management experience (ClearML / MLflow / DVC or equivalent)
• Proficient in using AI tools to accelerate development: Able to leverage AI coding tools to boost development efficiency while ensuring output quality

• Experience with multi-camera synchronization, Re-ID, MOT, or low-latency streaming (CMAF / LL-HLS / WHIP / WHEP)
• Experience with edge device operations (NVR / industrial PCs / PoE, thermal management, remote management / OTA) or building observability and rollback strategies
• Experience with retail / manufacturing / smart city field deployments, or building shared AI coding development environments for teams

1. Resume Screening
2. Online Interview
• Technical Initial Discussion (1 hour): Align on motivation, discuss past ML product deployment experience (architecture, trade-offs, outcomes)
3. On-site Interview
• Technical Deep Dive (1.5–2 hours): In-depth discussion of hands-on exercises, system design, and scenario-based questions
4. Final Interview
• Culture & Collaboration Interview (1 hour): Understanding collaboration style, communication approach, and conflict resolution
• CEO Interview (1 hour): Assessing personality traits and team culture fit
5. Offer & Onboarding

‍The entire process is expected to be completed within two weeks.