Are you ready to meet the demands of tomorrow? Evaluate your approach to skilling.

Deploy State-of-the-Art Vision Models from GPU Lab to Edge Device

12 Weeks. Live Online Classes. Instructor-led.

Enrol Now

Our Partners

Logos of NVIDIA, Google Cloud Platform, AWS, Databricks, and Azure.
A woman standing in front of a group of diverse people in a conference room, giving a presentation.

Push to edge GPUs or Jetson devices with KServe + KEDA autoscaling

Young woman wearing a virtual reality headset in a tech-filled environment, with digital interface visuals in front of her, suggesting virtual or augmented reality experience.

Train cutting-edge detection & segmentation models (YOLO-v8, Mask RCNN)

Group of five young adults collaborating around a computer in a modern office, engaged in software coding or programming.

Compress, optimise & serve with ONNX Runtime and NVIDIA Triton

What you will learn?

    • Best-practice labelling, augmentation & synthetic data generation

    • Reproduce YOLO-v8 detection baseline; evaluate mAP, FPS

    • Fine-tune Mask RCNN / DeepLab-v3+ on custom classes

    • Apply mixed-precision & gradient-accumulation to fit big models on modest GPUs

    • Export to ONNX, TensorRT, and OpenVINO; prune & quantise to INT8

    • Benchmark latency vs accuracy trade-offs; hit target FPS/Watts budget

    • Wrap models in NVIDIA Triton with gRPC; add FastAPI gateway and JWT auth

    • Deploy to KServe; build Helm charts for Jetson / Orin devices; configure KEDA for GPU auto-scaling

    • Capture inference metrics via Prometheus; create Grafana CV dashboards

    • Detect data drift with Evidently-CV; auto-trigger retraining pipelines in Argo Workflows

    • Team project: quality-inspection, safety-monitoring, or inventory-tracking solution

    • Deliver live edge demo, cost analysis, and 8-page design document to hiring-manager panel

Go beyond the notebook and put vision models where they matter: in the factory, on the drone, at the edge. Starting with dataset curation and YOLO-v8 fine-tuning, you’ll compress models with ONNX/TensorRT, serve them through NVIDIA Triton, and deploy to Jetson or KServe with auto-scaling GPUs. The capstone—an end-to-end detection or segmentation system—gives you portfolio proof for Computer-Vision or Edge-AI roles in manufacturing, mining, retail, and robotics.

Who Should Enrol?

  • You already train image models and now need to deploy object-detection or segmentation services that hit latency and GPU-cost targets.

  • Level-up from tabular models to large-scale visual data pipelines, augmentation strategies and active-learning loops.

  • Build perception stacks that run reliably on edge devices and stream telemetry to the cloud for re-training.

  • Learn how to host heavy vision models behind REST/GRPC gateways, scale them with autoschedulers and cache results.

  • Gain the vocabulary and architectural patterns to scope, cost and roadmap vision features for mobile, web or embedded products.

Engineers who must bring computer-vision models all the way to production hardware.

Prerequisites
Solid Python, basic deep-learning knowledge. Free Deep Learning Core + Docker-K8s Mini-Camp bridge badges are included for all Intermediate-track graduates.

Career Pathways

  • Design, train and deploy classification, detection and segmentation models at scale.

  • Optimise and quantise models for mobile, IoT and robotics hardware with TensorRT or ONNX Runtime.

  • Automate data versioning, model registry, CI/CD and real-time monitoring for high-throughput image pipelines.

  • Integrate perception models with control loops for drones, vehicles or industrial robots.

  • Scope end-to-end vision systems, balance cloud vs. edge costs, and ensure GDPR/ethics compliance.

Graduates leave with a fully-containerised vision service running on cloud & edge, a public GitHub repo, and recruiter-ready talking points that match the majority of “Computer-Vision Engineer” and “Edge-AI Engineer” roles advertised today.

Black and white photo of a smiling man with short hair and a beard, wearing a black Nike t-shirt, standing against a plain background.

Amir Charkhi
Technology leader | Adjunct Professor | Founder

With 20 + years across energy, mining, finance, and government, Amir turns real-world data problems into production AI. He specialises in MLOps, cloud data engineering, and Python, and now shares that know-how as founder of AI Tech Institute and adjunct professor at UWA, where he designs hands-on courses in machine learning and LLMs.

Advanced: Computer Vision @ Scale

12 Weeks. Live Online Classes.

Advanced: Computer Vision at Scale Advanced: Computer Vision at Scale
Quick View
Advanced: Computer Vision at Scale
$4,250.00

Go beyond the notebook and put vision models where they matter: in the factory, on the drone, at the edge. Starting with dataset curation and YOLO-v8 fine-tuning, you’ll compress models with ONNX/TensorRT, serve them through NVIDIA Triton, and deploy to Jetson or KServe with auto-scaling GPUs. The capstone—an end-to-end detection or segmentation system—gives you portfolio proof for Computer-Vision or Edge-AI roles in manufacturing, mining, retail, and robotics.

Student Outcomes

  • "I recently completed a course at AI Tech Institute and was thoroughly impressed. The curriculum was thoughtfully structured, the instructor was truly dedicated, and the overall learning experience delivered exceptional value."

    Maria

  • "Taking the Machine Learning course at AI Tech Institute was a game changer. It gave me the skills and confidence to break into AI and helped me land a job soon after. The content was practical and beginner friendly, with clear lessons and hands-on projects that made a real difference during interviews."

    Pani

  • "Very valuable course. Amir is a great instructor, explained how machine learning used in the real world in a easy way. Highly recommend for people who want to know more about AI or work in this area."

    Hoàng Phước

  • "The AI and Data science bootcamp was one of the most impactful learning experiences I’ve had. The curriculum was well structured. The instructor was knowledgeable and supportive, and the course provided a strong balance between theory and hands on labs. Highly recommended!"

    Ehsan

  • "I took a course at AI Tech Institute, and it genuinely exceeded my expectations. The content was not only clear and well-structured, but also incredibly relevant to the challenges I deal with day-to-day at work."

    Rosha

  • "Extremely happy with clear, well designed and applicable syllabus to be implemented in day-to-day need of nowadays coding ability. Coming from non-IT background, it is a valuable experience of seamlessly and effortless the teaching was, for me to understand the complex subjects. Would definitely recommend to colleagues and friends!"

    Kresna

  • "Absolutely fantastic experience with AI Tech Institute! I couldn’t be happier with the quality of education and support I received here. The instructors are true experts — knowledgeable, approachable, and always willing to help students succeed. The curriculum is up-to-date with the latest advancements in AI and tech, blending theory with plenty of hands-on projects that really build practical skills."

    Foad

  • "Amir is both technically proficient, and a very good teacher. He was attentive to where each person in the group was and paced the information flow accordingly. I also found him very approachable, encouraging and supportive."

    Francois

Frequently Asked Questions

  • Beginner courses: none— we start with Python basics.
    Intermediate & Advanced: ability to write simple Python scripts and use Git is expected.

  • Plan on 8–10 hours: 2× 3-hour live sessions and 2–4 hours of project work. Advanced tracks may require up to 10 hours for capstone milestones.

  • All sessions are recorded and posted within 12 hours. You’ll still have access to Slack/Discord to ask instructors questions.

  • New intakes launch roughly every 8 weeks. Each course page shows the exact start date and the “Apply-by” deadline.

  • Just a laptop with Chrome/Firefox and a stable internet connection. All coding happens in cloud JupyterLab or VS Code Dev Containers—no local installs.

  • Yes. 100 % refund until the end of Week 2—no questions asked. After that, pro-rata refunds apply if you need to withdraw for documented reasons.

  • Absolutely. We issue invoices to companies and offer interest-free 3- or 6-month payment plans.

  • Live Q&A in every session, 24-hour Slack response time from instructors, weekly office-hours, and code reviews on your GitHub pull requests.