Data Scientist - Computer Vision

Job Summary Qualification & Experience

• Education: Master's or Bachelor's degree in Computer Science, Data Science, Electrical Engineering, Mathematics, or a related quantitative field.
• Experience: 5+ years of hands-on experience in Data Science and Machine Learning, with at least 3+ years dedicated specifically to solving Computer Vision applications in production settings.

Key Responsibilities

• End-to-End Model Lifecycle Development: Design, build, train, evaluate, and optimize custom Computer Vision models (classification, object detection, segmentation, tracking, OCR, visual inspection) from concept to production.
• Data Pipeline & Annotation Strategy: Establish data collection, cleaning, and labeling pipelines; evaluate and leverage annotation platforms (e.g., CVAT, Labelbox); define guidelines to ensure high-quality training datasets.
• Business Problem Translation: Partner directly with business leaders and product teams to translate ambiguous business requirements into practical, well-scoped Computer Vision problems with clear KPIs.
• Cloud Architecture & Deployment: Architect scalable ML pipelines on cloud platforms (AWS or GCP) using containerization and serverless/managed machine learning services.
• Model Optimization & MLOps: Quantize, compress, and optimize models (e.g., using ONNX, TensorRT, OpenVINO) for low-latency inference on cloud or edge environments; set up monitoring for model drift and performance.
• Technical Leadership & Mentorship: Guide full-stack/ML engineers on best practices in model design, code quality, research methodology, and experimentation tracking.
• Continuous Innovation: Stay up-to-date with recent advancements in Computer Vision, Vision-Language Models (VLMs), and modern AI architectures to evaluate buy-vs-build options and bring novel ideas to the team.

Technical Essentials

• Deep Learning & Frameworks: Expert proficiency in Python and core deep learning frameworks ( PyTorch or TensorFlow/Keras ).
• Computer Vision Ecosystem: Camera fundamentals, basics of image and video encoding, working knowledge of Camera calibration, 3d reconstruction and Multi camera multi object tracking. Experience using CV tools and frameworks like OpenCV , Nvidia Deepstream, Colmap and DL architectures for CV tasks like object localization, feature extraction and matching and foundation models.
• Cloud Proficiency (AWS or GCP):
  • AWS Stack:
    SageMaker, S3, EC2, Lambda, Rekognition, ECR/EKS.
  • GCP Stack:
    Vertex AI, Google Cloud Storage, Cloud Run, Vision API, Compute Engine.
• Data Engineering & Tooling: Hands-on experience with dataset versioning (e.g., DVC), annotation management, and relational/NoSQL databases.
• Software Engineering & MLOps: Proficiency in standard software practices Git, unit testing, modular coding, Docker, REST API design (FastAPI/Flask), and ML experiment tracking tools (e.g., MLflow, Weights & Biases).

Soft Skills & Business Essentials

• Business Acumen: Ability to link technical metrics (e.g., mAP, IoU, F1-score) directly to business outcomes (e.g., operational efficiency, cost reduction, accuracy thresholds).
• Stakeholder Management: Outstanding verbal and written communication skills to present technical findings clearly to non-technical business leaders.
• Problem-Solving Mindset: A structured, analytical approach to troubleshooting complex edge cases in unstructured visual data.

Nice-to-Have / Added Advantages

• Professional AWS (e.g., AWS Certified Machine Learning - Specialty) or GCP (e.g., Google Professional Machine Learning Engineer) certifications.
• Experience with Generative AI for Vision (Diffusion Models, Vision-Language Models, zero-shot detection).
• Experience deploying models to Edge Devices (NVIDIA Jetson, Raspberry Pi, Android/iOS with TFLite/CoreML).
• Track record of published research papers (IEEE, CVPR, ICCV, ECCV) or top-tier Kaggle computer vision achievements.

Success Metrics

• Production Deployment: Timely delivery of robust, high-accuracy CV models into production

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.