active builder & slop hater, interests in AI/ML, GPUs, systems programming, building automation.

Applied ML, MLOps pipelines, RAG/Conversational search & AI tooling, Attention mechanisms in CNNs, BERT-based long-term prediction (NLP), LSTMs for building energy analytics, Computer vision (C++ virtual try-on, number plate recognition), Lighting/IoT systems, Open-source ML tooling, Research-to-product ML engineering

currently looking at writing my own custom inference runtime to learn more about GPUs :)

I am a Data Scientist at Saal, with over 6 years of experience in data science and analytics. With a Master's degree in Applied Mathematics from Stellenbosch University and a Bachelor's degree in Computer and Information Systems from Sudan University of Science and Technology, I have developed machine learning and deep learning models to analyze diverse datasets, including time-series and images. I have previously worked at DP World, Sila and Ether Technologies. I am an active member of the data science community, I volunteer at Zindi and previously at the Deep Learning Indaba. Last year, I have started a YouTube channel to share my knowledge in the field with everyone who is interested in Data Science and find a language barrier, hence, all the content is shared in Arabic.

I am interested in the applications of machine learning in several domains including social data, education, and psychology.

Currently working at Saal as a Data Scientist, developing deep learning models for analyzing time-series datasets. Projects involve applying data science techniques scrape, clean, and analyze data. I am very much interested in the applications of data science in different domains including aviations, social science, and education.

AI Tinkerer, realist, borderline AI doomer but write AI apps for a living :)

ML, AI, High performance computing, GOLANG!!!

Currently working on a couple of very interesting (but also confidential) work problems. Mostly involving large scale and heavily parallelized computer vision deployments. Smart systems that can be used to detect incidents on the fly from a large number of livestreams ingesting 100s of frames a second while using LLM's in combination with some smaller more efficient models to detect incidents and annotate them. Also working on other systems to produce annotated live streams with little delay!

I'm a Software Engineer at Flow71 focused on building production-ready AI systems, automation tools, and rapid prototypes. My work spans LLM-powered applications, workflow automation, computer vision, and data-driven systems. I enjoy shipping fast, experimenting with new ideas, and collaborating with builders who are pushing practical AI into real-world use.

LLM-powered systems, AI agents, workflow automation, and applied AI in real-world environments. Interested in collaborating on tools that improve productivity, education, and decision-making at scale.

Production AI systems and rapid prototypes including: • Clam.run - AI agent that lives in LinkedIn posts and comments • OrgFlow - AI-powered org chart builder generating full structures from minimal input • RightQ.ai - AI interview assistant generating domain-specific questions instantly • Deepfake Detection - ResNet-based model achieving 93%+ accuracy on deepfakes