Description :

Were building an AI-powered document scanning app and need a Senior Computer Vision/OpenCV Engineer who can create world-class document enhancement filters similar to CamScanner Magic Pro, Adobe Scan Filters etc

If you love turning low-quality images into clean, professional documents using a mix of CV + lightweight deep learning, this role is for you.

We need to build following models first :

1. Document detection model under 10ms (inference on android)

2. Document segmentation model under 10ms

3. Document dewarp model under 50ms

4. Appearance enhancement model under 50ms (deshadowing, denoising, deblurring, brightness + contrast balancing, color preservation, binarization, text clarity enhancement etc)

5. Then finger removal, watermark removal, handwritten text removal, sign removal, seal removal model, OCR enhancement model etc in the future

6. Then document naming auto suggestion, document summary, Q&A support etc gen AI features

You will own the entire document enhancement pipeline used by millions of users

Key Responsibilities :

- Design and implement ultra-fast document detection models with sub-10ms inference on Android.

- Build document segmentation models optimized for real-time mobile performance.

- Develop document dewarping algorithms with high geometric accuracy under strict latency constraints.

- Create advanced appearance enhancement pipelines including deshadowing, denoising, deblurring, brightness and contrast balancing, color preservation, and binarization.

- Optimize text clarity and readability for OCR-ready document outputs.

- Deploy and optimize CV and ML models using TensorFlow Lite for on-device inference.

- Continuously benchmark and improve output quality against industry-leading scanning apps.

- Collaborate with Android and product teams to integrate models into production pipelines.

- Own end-to-end model lifecycle from experimentation to production release.

- Research and prototype future features including finger removal, watermark removal, handwriting removal, seal removal, OCR enhancement, and GenAI-based document intelligence.

Requirements :

- 3+ years in Computer Vision, Image Processing, OpenCV

- Proven experience building and deploying TFLite models on mobile

- Strong understanding of model optimization for real-time inference

- Ability to reproduce or exceed filters from CamScanner / Adobe Scan

- Strong engineering discipline: clean code, structure, documentation

Bonus skills :

- Android app integration experience (Java/Kotlin + native CV/ML)

- Experience with hybrid CV + CNN filter design

- Knowledge of GANs or any advanced enhancement models

What Were Looking For :

You will build document enhancement filters that outperform CamScanner Magic Pro in clarity, contrast, readability, color accuracy, real-time performance. Our benchmark is world-class output, not academic prototypes

About Us :

We build next-gen AI products designed to simplify and enhance everyday experiences. Our mission is to apply the best of computer vision, deep learning, and NLP to create intelligent tools for massive real-world impact

Why Join Us?

- Be part of a fast-moving AI product team

- Work on one of the hardest real-world CV problems

- Solve real-world visual problems used by 800K+ users

- Ship models used by billions of users

- Remote-first team with flexible work culture