A

Almetra

Edge AI vision analytics for manufacturing

Computer Vision Engineer - Video & VLM

Computer VisionUpdated on Nov 19
Full-timeBerlin, GermanyHybrid

About the Job

Requirements

  • Experience building computer vision and deep learning models for video, ideally deployed for real users.
  • Hands-on experience fine-tuning vision-language models or large video models such as InternVL or Qwen-VL.
  • Strong PyTorch and multi-GPU training experience, with clean and reproducible Python.
  • Experience running models in production pipelines in the cloud and/or on edge devices, using tools such as Docker, AWS or Jetson.

Responsibilities

  • Improve the computer vision pipeline end to end, from person and hand tracking to action recognition with vision-language models.
  • Fine-tune and evaluate vision-language models on real factory video, especially on steps that look alike.
  • Run models reliably on video in the cloud and on factory edge devices, with checks that catch weak results early.
  • Make models faster and cheaper to run through distillation, quantization and smaller backbones.
  • Help shape the company's approach to video understanding.

Desired Qualifications

  • Experience with person detection and tracking in crowded scenes, especially transformer-based detectors such as the DETR family and Grounding DINO.
  • Experience with TensorRT, quantization and inference optimization for edge devices.
  • Experience with temporal grounding or video captioning.
  • Publications on video or multimodal learning.
  • Experience with industrial, retail or other real-world camera settings.

About the Company

Almetra

Edge AI vision analytics for manufacturing

Almetra equips production lines with AI cameras to convert shop-floor video into real-time data, including cycle times, output, utilization, and micro-stoppages. Processing happens at the edge with privacy-by-design anonymisation, and the system links vision analytics with MES/PLC/ERP context to produce actionable insights rather than just dashboards. By combining line-level analytics with input from continuous-improvement engineers, it helps factories diagnose root causes of productivity losses and drive improvements. The goal is to scale the product and expand into the US, supported by funding such as the June 2026 Series A led by blisce/.

Company size
51-200
Company stage
Series A
Total funding
$22.4M
Headquarters
Berlin, Germany
Founded
2022

Recently Posted Jobs

Explore more opportunities at Almetra

See more jobs