Open to opportunities

Hakim Ali

@hakimali1

Message

I build and deploy edge computer-vision systems and explain advanced AI research visually.

Pakistan

Message

What I'm looking for

I'm looking to build learning systems that move beyond correlation toward causal reasoning, world models, and embodied decision-making, especially in computer vision and reinforcement learning. I'm open to relocation.

At Motive, I own end-to-end delivery of temporal computer-vision models for safety and event detection, from data curation and architecture design through large-scale training and on-device deployment. I shipped quantization-aware object-detection models for real-time, memory- and latency-constrained edge hardware and applied reinforcement learning to improve precision.

I also founded AGI Lambda, where I code animated Python and Manim explanations of deep learning research for a 110,000+ audience across YouTube and Instagram. I cover reinforcement learning, world models, JEPA, YOLO, Vision Transformers, LSTMs, and Grad-CAM while running a global Discord community of 350+ AI researchers and practitioners.

My work spans computer vision, multimodal learning, NLP, explainable AI, and multi-agent reinforcement learning, including a top-5 SemEval-2024 result for multilingual persuasion-technique detection in memes.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer (L3)

Current

Motive

Jul 2025 - Present (1 year 2 months)

Owns end-to-end delivery of temporal computer-vision models for safety/event detection, including data curation, architecture design, large-scale training, evaluation, and on-device deployment. Shipped quantization-aware object-detection models that reduced model size and inference latency while maintaining accuracy, enabling real-time detection on edge hardware.