Senior Machine Learning Engineer

A highly skilled Senior Research Scientist with extensive experience in deep learning, computer vision, and AI-driven solutions. Proven track record of developing cutting-edge technologies, including AI systems for image categorization, object detection, and event detection. Adept at driving innovation through multi-task neural networks, vision language models, and efficient image annotation tools. Experienced in leading cross-functional teams and designing next-generation hardware and software solutions. Demonstrates a strong ability to reduce training costs, improve system accuracy, and contribute to large-scale deployments globally.

‍

Senior Research Scientist (2018 - Present)

‍
Winnow Solutions, London, UK

‍
• Co-patented a vision AI system for categorizing food-related images.
• Designed multi-task hybrid deep neural networks, reducing the amount of required training data and shortening the time from data collection to model deployment.
• Developed object detection networks with 98% accuracy, working reliably in diverse environments globally.
• Invented memory-efficient event detection neural networks for video sequences, achieving up to 100 fps with a low error rate.
• Proposed wave filtering algorithms to stabilize weight-measuring systems, benefiting large-scale customers.
• Developed specialized image annotation tools for segmentation, achieving high throughput and saving substantial costs.
• Led initiatives to design and test next-gen cameras and lighting for improved recognition accuracy.
• Developed early object prediction capabilities for next-gen products via a multi-task deep net, achieving high precision and recall in early-phase testing.
• Developing proprietary vision language models (VLM) for food-related image analysis.

‍

Tech Chef & Co-founder (2017 - 2018)

‍
Entrepreneur First, London, UK

‍
• Built big data processing pipelines for operational satellite imagery with national-scale capabilities.
• Developed predictive modeling techniques for forecasting, achieving 80-90% accuracy using multi-modal feature engineering.
• Worked with insurance brokers to specialize products for specific markets.

‍

Research Engineer (2014 - 2016)

‍
Lab for Integration of Systems and Technology (CEA), Paris, France

‍
• Leveraged weakly supervised learning to train convolutional networks on large-scale noisy image data, achieving state-of-the-art performance.

‍

Research Assistant (2010 - 2012)

‍
Research Institute, Paris, France

‍
• Developed graph-based dimensionality reduction techniques for image analysis.

‍

Research Assistant (Internship) (2009)

‍
National Institute of Informatics, Tokyo, Japan

‍
• Developed space-time feature descriptors for human event detection.

‍

Qualifications

‍
• PhD in Signal & Image Processing
• Masters in Computer Science
• Bachelor's in Information Technology

‍

Domains of Expertise

‍
• Deep Learning
• Convolutional Neural Networks
• Computer Vision
• Large Vision Language Models
• Efficient Computing
• Information Retrieval
• Multi-modal Signals

‍

Soft Skills

‍
• Project Leadership
• Team Collaboration
• Planning and Execution
• Problem Solving
• Communication

‍

Skills

‍
• Python, C, C++, SQL
• TensorFlow, PyTorch
• ONNX, TensorRT, Nvidia TX2, RPi, Rockchip
• AWS, GCloud, Docker

‍

Awards

‍
• Patent for Vision AI System
• Startup Grant by Entrepreneur First
• Talent Recognition by TechNation
• Best Paper at IGARSS

‍

Publications

‍
• Over 20 publications in top conferences and journals, including CVIU, SIGIR, BMVC, ICIP
• PhD Thesis: "Kernel Learning and Applications"

‍