Relevant Experience & Education Highlights

Delivers 10 years of hands-on expertise in computer vision and machine learning, with a proven track record of designing, optimizing, and deploying production models for surveillance, industrial monitoring, and smart city applications.

1. Design, develop, and deploy advanced computer vision models for real-world applications, including object detection, tracking, OCR, image search, and scene understanding

Excelled in developing object detection, multi-view analysis, and surveillance systems tailored for industrial and urban environments.

• Developed holistic multi-view camera systems combining inputs for novel ergonomic risk scoring at a global logistics company, managing end-to-end from proof-of-concept to production deployment.
• Engineered custom YOLO-based detection alongside segmentation and cognition pipelines for sequence understanding in autonomous systems at an innovative propulsion startup.
• Architected scalable surveillance solutions for outdoor and infrared imagery with instance and action detection at a leading computer vision research center.
• Led CNN model development for intelligent video analytics in smart city applications at a specialized analytics firm.

2. Build and optimize deep learning models, ensuring high accuracy, performance, and scalability for deployment in production environments

Achieved substantial performance improvements through refactoring, optimization techniques, and hardware acceleration.

• Refactored production code boosting throughput by 57% via OOP principles, redundancy reduction, and ONNX-optimized multi-processing pipelines at a global logistics company.
• Redeployed human object detection models reducing false positives and improving mean average precision by 10%, minimizing factory downtime at a global logistics company.
• Spearheaded Deepstream framework adoption with Nvidia GStreamer for GPU-optimized video analytics scaling to large deployments at a specialized analytics firm.
• Benchmarked inference pipelines yielding 1.5-2x runtime improvements on real test imagery using ONNX at a global logistics company.

3. Lead the design and implementation of scalable pipelines for data processing, model training, and model deployment

Architected end-to-end pipelines and monitoring systems enhancing operational efficiency and developer productivity.

• Designed proof-of-concept camera systems with software integration, APIs, and server pipelines for sports tech at a stealth startup.
• Consolidated inference reporting scripts cutting redundant operations by 60-80% and saving engineering time at a global logistics company.
• Implemented log server monitoring with email alerts for uptime/downtime, reducing manual checks by 5-8 minutes daily at a global logistics company.
• Directed pipeline expansions for server integration in sports imaging systems at a stealth startup.

4. Optimize models for performance on various hardware platforms, including CPUs, GPUs, and edge devices

Optimized models leveraging GPU acceleration, edge frameworks, and runtime enhancements for diverse deployment targets.

• Utilized ONNX runtime optimizations in multi-processing pipelines for 1.5-2x inference speedups across CPU and GPU benchmarks.
• Integrated Nvidia Deepstream and GStreamer for efficient GPU utilization in large-scale video analytics deployments.
• Developed systems with RPI and Sony IMX cameras including low-level controls for edge-based sports tech proof-of-concept.

Requirements & Candidate Alignment