Computer Vision & Deep Learning Scientist — Document Verification & OCR (Identity)

Our mission is to create innovative, robust, and user-friendly digital identity solutions. We are looking for a passionate and skilled Computer Vision & Deep Learning AI Scientist specializing in identity document verification and OCR to join our dynamic team. Your work will directly impact the trustworthiness, accuracy, and inclusiveness of VIDA's identity verification pipeline — the layer that decides whether a person's ID document is genuine, readable, and matches the data we hold.

Responsibilities:

• Design, train, and deploy deep learning models that verify whether an identity document is authentic or fraudulent: detecting photo substitution, digital tampering (copy–paste, splicing), inconsistencies in fonts and templates, photographs of screens, printed reproductions, and fully AI-generated documents.
• Build production-grade OCR and field extraction pipelines that read and structure information from national ID cards, passports, driver's licenses, residence permits, family cards, tax IDs, utility bills, and bank statements.
• Own document image quality end-to-end: detecting blur, glare, skew, and occlusion at capture time; correcting perspective and distortion; cropping; and giving the user clear, real-time guidance on how to retake.
• Own datasets end-to-end: collection strategy across geographies and document templates, labeling specifications, quality gates, hard-negative mining, and fairness audits across document age, wear, lighting, and country of issue.
• Build realistic benchmarks and red-team suites of tampered documents, recaptured images, and synthetic IDs; report meaningful metrics (true-positive rate at low false-positive rate, field-level accuracy, character and word error rates, end-to-end pass-through rate) and run honest ablations and stress tests.
• Ship models with clear latency and accuracy targets; partner with engineers on scalable inference, on-device variants where required, and post-deployment monitoring for drift, false-positive hot spots, and new attack signatures.
• Translate emerging fraud patterns — template farms, AI-generated selfies paired with synthetic IDs, screen-recapture attacks, composite "Frankenstein" IDs — into model features and detection rules; partner with product, fraud ops, compliance, and customers in banking, fintech, and telco.
• Track the state of the art in document AI and adopt techniques that improve accuracy, security, latency, or cost.
• Work independently or in a team to solve complex problem statements.

Requirements:

• Advanced degree in a quantitative field or equivalent experience, and 3+ years building and launching computer vision and deep learning models (document verification, OCR, or document forensics strongly preferred).
• Strong foundations in modern computer vision and deep learning: detection, segmentation, sequence models for text recognition, self-supervised and contrastive learning, and techniques for making models robust to domain shift.
• Hands-on experience in at least one area of document AI: text detection and recognition, document layout understanding, or document tampering forensics.
• Proficient in Python and PyTorch (or TensorFlow); strong working knowledge of OpenCV and image processing libraries.
• Proven track record taking models from notebook to production, with experiment tracking, CI/CD, and post-deploy monitoring.
• Clear communication, ownership mindset, and bias to ship.

Nice to have:

• Experience with document forensics: detecting compression artifacts, manipulation traces, and artifacts from GAN or diffusion-generated images.
• Experience with machine-readable travel documents (eMRTDs): machine-readable zones, NFC chip data, and barcode decoding.
• Multilingual and low-resource OCR, scripts beyond Latin (Thai, Chinese), handwriting recognition, and Indonesian document templates (KTP, SIM, Paspor, NPWP, KK).
• Building synthetic data and document generators for both training and red-teaming.
• On-device and low-latency inference (TensorRT, ONNX Runtime, TFLite, Core ML), quantization, and distillation — for mobile capture flows.
• Familiarity with cloud platforms (AWS, GCP, Azure) for model deployment.
• Familiarity with relevant standards and frameworks: ICAO 9303 (passports and machine-readable travel documents), ISO/IEC 18013-5 (mobile driver's licence), and NIST SP 800-63A (identity proofing).

What are we trying to solve? We have 7.5 billion people on Earth, of which over 1 billion cannot securely prove their identity right now. Every year, 140 million babies are born, of which 40 million go unregistered. Simply put, these people are deprived of social benefits, such as education and health, their civil rights to vote and travel; and are excluded from the economy because they cannot sign up for bank accounts, loans, welfare programs etc. We believe this is unacceptable, and needs to change.

At VIDA, we are creating a frictionless digital identity system. One that fulfils the needs and expectations of our times, and is available anywhere, for everyone.

Why are we solving this problem?

The United Nations (UN) and World Bank ID4D initiatives aim to provide everyone on the planet with a legal identity by 2030. This deadline is just a few years away, we are expecting a digital identity to be a legal human right by then and we at VIDA want to be pioneers in leading this change.

Who are we?

We are a highly driven bunch of people solving this problem for our own reasons. Whether it is misleading doctors, or because we didn't get access to fair ration due to corruption — our collective goal aligns.