Middle/Senior Computer Vision and Machine Learning Engineer

🟠 Details: we are looking for a Middle/Senior Computer Vision & Machine Learning Engineer to join our team and work on innovative AI and Computer Vision solutions for international clients.

One of the projects you may join focuses on developing an AI-powered system for extracting structured spatial data from 2D and 3D floor plan imagery. It is a practical, R&D-driven project involving room detection, layout extraction, area estimation, and solving real-world computer vision challenges such as perspective distortion, occlusion, and scale estimation.

🟠 Qualifications:
3+ years of hands-on experience in Computer Vision & Machine Learning.
Strong practical experience with image segmentation, object detection, image preprocessing, and geometry-based computer vision.
Experience with OpenCV and classical computer vision techniques.
Experience with modern segmentation and detection architectures such as YOLO, U-Net, SegFormer, Detectron2, SAM, or similar.
Strong Python engineering skills.
Experience with polygon/vector processing tools such as Shapely, OpenCV contours, or similar geometry libraries.
Ability to work with imperfect visual inputs and design robust fallback logic.
Understanding of OCR-based extraction using tools such as Tesseract, PaddleOCR, docTR, or cloud OCR services.
Ability to design end-to-end CV pipelines from experimentation to reproducible prototypes.

🟠 Would be a plus:
Experience with floor plan recognition, architectural drawings, construction tech, real estate tech, or takeoff/estimation workflows.
Experience with perspective correction, homography, depth estimation, vanishing lines, or top-down reconstruction.
Experience with depth models such as MiDaS or Depth Anything.
Experience with synthetic data generation or image-to-image normalization workflows.
Experience building ML evaluation and observability pipelines.
Experience integrating ML pipelines into backend services or APIs.
Experience leading technical direction for R&D-heavy projects.

🟠 Responsibilities:
Design and implement a render-to-plan pipeline for converting 3D floor plan renderings into simplified 2D measurement-ready layouts.
Develop CV/ML approaches for room segmentation, boundary detection, polygon extraction, room classification, and space labeling.
Work with complex visual inputs including top-down plans, oblique renders, dollhouse-style layouts, and low-quality marketing imagery.
Estimate room-level areas using geometry, OCR, scale evidence, metadata, and heuristic-based validation approaches.
Define evaluation logic, confidence scoring, and quality metrics for pipeline outputs.
Produce structured outputs and visual QA overlays for downstream flooring takeoff workflows.
Collaborate with AI/backend engineers to integrate the CV pipeline into a web-based PoC tool.
Document technical assumptions, limitations, risks, tradeoffs, and fallback strategies.

🟠 Interview steps:
Recruiter screening (about 45 minutes).
Take-home task.
Technical ML/CV interview (about 1 hour).
Live coding session (about 1 hour).