Open to opportunities

Moazam Ali

@moazamali

Message

Senior Deep Learning Engineer specializing in computer vision, multimodal MLLMs, and green AI.

Pakistan

Message

What I'm looking for

I’m looking for a role where I can ship production computer-vision and multimodal MLLM systems (RAG, grounding, anomaly detection) with strong engineering rigor, while prioritizing green AI/carbon optimization and real-world measurable outcomes.

I’m a Senior Deep Learning Engineer with 4+ years shipping production computer-vision, multimodal MLLM, RAG, and anomaly detection systems for streaming, transmission inspection, and remote sensing. I focus on measurable impact—like delivering 3D reconstruction from 2D imagery for precise measurements and building multi-GPU training pipelines under extremely limited labels.

I’ve developed RAG-based multimodal MLLM solutions for real-time, zero-shot labeling and guidance, fine-tuned Qwen 3.5-VL for defect grounding and detection, and reduced false positives in abnormal login detection by >25% under data drift. I’m also passionate about green AI/carbon-footprint optimization and currently pursue my Master’s in Artificial Intelligence while working remotely full-time for a U.S. client.

Experience

Work history, roles, and key accomplishments

Current

Deep Learning Specialist

Current

Meriti Inc

Jul 2024 - Present (1 year 9 months)

Led computer-vision pipelines for transmission infrastructure inspection on 100-MP drone imagery, including 3D reconstruction from 2D (MoGe/Metric Depth) and multi-GPU training with on-the-fly augmentation under limited labels. Built a RAG-based multimodal MLLM for inspector guidance and fine-tuned Qwen 3.5-VL for real-time grounding; also reduced false positives by >25% in hybrid temporal CNN+Tra

3D Reconstruction RAG Qwen 3.5 VL Fine Tuning Multi GPU Training Anomaly Detection Computer Vision Vision Transformer Green AI Optimization

Machine Vision Engineer

HCRL, UET

Mar 2023 - Jun 2024 (1 year 3 months)

Designed and deployed production AI vision systems for industrial safety and retail analytics, including PPE compliance monitoring with 87% violation detection and mAP 0.93, and retail analytics using YOLOv8 object tracking with heatmaps achieving F1 0.97 on a 5,000-image dataset.

YOLOv8 Heat Mapping Transfer Learning Computer Vision OpenCV PyTorch Model Evaluation