Freelance · Available for new projects

Computer vision, from prototype to production.

Freelance computer vision engineering for startups and product teams: custom models, video pipelines, and the deployment work that gets them into a real product.

8+ yrs in industry Research → production Fixed-scope projects

Illustration: a street scene with people, cars and a dog, each outlined by a labelled detection box and a tracking trail.

Object detectionSegmentationMulti-object tracking Open-vocabulary detectionVideo searchSynthetic data Generative AIModel optimisationCloud & edge deployment

Selected projects

A few systems that are live today, covering video search, real-time detection and synthetic training data.

Vantage Search interface showing a grid of matching video frames

Case 01 · Video understanding

Vantage Search

Natural-language search across video libraries.

A video indexing and semantic search platform. Videos are indexed so they can be queried in plain language, returning the moments that match.

  • Video indexing
  • Semantic search
  • Multimodal AI
Flash-Findr dashboard with live camera feeds and detections

Case 02 · Real-time video

Flash-Findr

Real-time object detection and tracking in the browser.

A deployed web application for live video analysis, running detection and tracking end to end, from model inference to the user interface.

  • Real-time
  • Detection
  • Tracking
  • Cloud Run

Case 03 · Synthetic data

Dataspan.ai Platform

Synthetic training data for visual inspection.

A platform for generating AI training datasets and training models on them. The demo shows synthetic "pitting" defects generated on metal threads, for training inspection models where real defect images are scarce.

  • Generative AI
  • Synthetic data
  • Visual inspection

● In progress

VisionPilot, a tool that builds and runs vision pipelines from a plain-language description. Demo coming soon.

How I can help

Projects usually cover one or more of these stages, depending on where you are.

01Scope

Feasibility study

Assessing whether and how the problem can be solved: reviewing relevant research, prototyping on your data, and outlining a plan with an estimated cost.

02Build

Models & pipelines

Custom models for detection, segmentation, tracking, classification and generative tasks, trained on your data and tuned to your performance and hardware requirements.

03Ship

Deployment

Deployment to cloud or edge, model export and optimisation, monitoring, and documentation so your team can maintain the system.

Pricing

Projects are fixed-scope and fixed-fee. Deliverables, timeline and cost are agreed before work begins.

Portrait of Almog Elharar

Background

I'm Almog, a computer vision engineer with 8+ years of industry experience building machine learning systems, with a focus on bringing research methods into reliable production software.

My work covers detection, tracking, video understanding and generative models for synthetic data. On freelance projects I work directly with your team throughout, from the initial research to deployment.

Focus
Computer vision · Video · GenAI
Toolkit
PyTorch · TensorFlow · ONNX
Elsewhere
GitHub · LinkedIn

Have a project in mind? Let's talk it through.

  • Free intro call
  • Written plan and fixed quote