Skip to content
freelancermapOpen

Computer Vision / Applied ML Engineer– Smart Retail, Remote

ForTech Consulting GmbH · Frankfurt am Main, Deutschland, Germany · Remote · Senior

  • Frankfurt am Main, Deutschland, Germany · Remote
  • 12 Monate
  • English

Indexed (today) · refreshed constantly

Start
2026-12
Duration
12 Monate
Work mode
Remote
Contract
Freelance / B2B
Seniority
Senior
Reference
3076348

Required skills

Nice to have:Android DevelopmentKotlinJavaC++JNICamera ProcessingONNXModel QuantizationQualcomm QAIRTMachine Learning

Languages: English

Job description

Project overview

We are looking for an experienced Computer Vision / Applied ML Engineer for a long-term project in smart retail. Our client is building camera-based verification for smart shopping trolleys: the system confirms whether the product a customer scanned was actually placed in the basket.

You will work with real shopping videos to detect placement events, track products, and distinguish store products from personal belongings. You will also develop a cloud-based data preparation and training pipeline for lightweight models that will run on Android/Qualcomm devices.

You will own your area of the project and work closely with the project lead. The client’s development team is available for questions at any time.

Responsibilities

Required skills

  • Review existing experiments and technical plans and take them through implementation, evaluation, and deployment-ready model exports
  • Build maintainable, repeatable pipelines rather than research notebooks or one-off demos
  • Set measurable accuracy and performance benchmarks, especially for correct-product confirmation versus incorrect acceptance
  • Make evidence-based model and architecture decisions that balance accuracy, device limitations, and training/inference costs
  • Strong hands-on experience with Python, PyTorch, and OpenCV
  • Practical experience training and fine-tuning object detectors such as YOLO
  • Experience with video analysis, object tracking, image embeddings, and metric learning for product recognition
  • Ability to build reliable datasets: annotation workflows, automated labeling, data curation, and quality control
  • Experience evaluating models on real-world data with occlusion, motion blur, changing lighting, and few examples per product
  • Comfortable with Linux, Docker, GPU training, and reproducible experiments

Nice to have

Interested?

Please send your CV in English with relevant project references, your availability, and your hourly rate.

  • Android development with Kotlin/Java and native C++/JNI, plus camera and video-processing experience
  • Edge-model export, quantization, and optimization with ONNX and Qualcomm QAIRT / Hexagon DSP / GPU tooling

Project details

15.12.2026 or 02.01.2027

Long-term

30–40 hours per week

100 % remote

Language: English

  • Start: 2026-12
  • Duration: 12 Monate
  • Remote: remote
  • Location: Frankfurt am Main, Deutschland, DE
  • Languages: English

Indexed by 9am's meta-search. You are a job provider? Claim or manage this listing

Found on FreelancerMap DE

  1. Job boardNew

    Financial Risk Management Expert

    Bosshard & Partner Unternehmensberatung AG· Wallisellen / Remote (nur aus der Schweiz), Switzerland· Hybrid· Senior

    • Financial Risk Management
    • Credit Risk
    • Market Risk
    • +5

Browse all jobs
Sign up free to apply