Bestoun S. Ahmed

Proud to share our new paper in JSS! 🎉 An MLOps + AI-driven digital twin to make steel production smarter, greener & more efficient. 🔗 doi.org/10.1016/j.jss.… #SmartManufacturing #MLOps #AI

🎯 Model-Based Testing gets AI upgrade! 2025 brings automated test scenario generation from user stories. Smart coverage that adapts in real-time 🧠✨ #ModelBasedTesting #TestGeneration #QualityEngineering #AI #testing #TestAutomation

⬅️ Shift-Left testing is the new QA superpower! Catch bugs early when they cost pennies, not millions. Early detection = faster delivery = happier devs 🔧 #ShiftLeft #QualityEngineering #DevOps #Testing #TechTrends2025

Tweet 5:💊 AI is CRUSHING drug discovery timelines! Machine learning models predict compound efficacy, slashing years off development. From disease markers in genomics to literature analysis—AI is becoming the ultimate research accelerator. 🧬 #AIHealthcare #DrugDiscovery

Tweet 4:🌍 "World Models" trained on massive video datasets are learning REAL-WORLD PHYSICS! This could be the biggest AI breakthrough since ChatGPT. Imagine AI that truly understands how the physical world works. Game. Changer. 🚀 #WorldModels #AI #Physics #MachineLearning

I've identified the most engaging and tweetable insights from the latest AI research. Here are 15 carefully crafted tweet drafts that balance technical depth with accessibility:

Proud to have supervised Firas Bayram, who successfully defended his PhD on April 25 🎓His work on adaptive ML systems bridges the gap between research and real-world deployment. Thanks to the examination committee and opponents for their valuable contributions #PhD #MLOps

Happening now at ITEQS'25: Firas is presenting their paper on "Quality Assurance for LLM-RAG Systems: Empirical Insights from Tourism Application Testing."

Towards Trustworthy Machine Learning in Production: An Overview of the Robustness in MLOps Approach. arxiv.org/abs/2410.21346

🎉 Our paper on Trustworthy ML in Production has been accepted in ACM Computing Surveys! With @Firas__P , we provide a comprehensive overview of robustness in MLOps - from automation to DataOps and ModelOps. Read it here: arxiv.org/abs/2410.21346 #MLOps #AI #TrustworthyAI

@shukat_ali Indeed

@bestoon82 We can replace editor with LLM!

In the era of AI and rapid development of technology and research, letting an article be under review for 2.5 years is really not acceptable; that means, as an editor, you are not doing your job. #Journal

🚀 Researchers are making AI faster and more reliable by "shrinking" LLMs through smart compression techniques. This boosts speed and enhances fairness and trustworthiness in AI decisions. Learn more about this exciting advancement! arxiv.org/abs/2401.05561 #AI #Tech #Innovation

New research in Nature reveals a surprising AI pitfall: training models on AI-generated data can cause "model collapse," leading to nonsense outputs over time. nature.com/articles/d4158… #AIResearch #ML #DataScience #ArtificialIntelligence #GenerativeAI

🎉 Excited to share our paper, "Adaptive Data Quality Scoring Operations Framework using Drift-Aware Mechanism for Industrial Applications," is accepted in @JSSoftware🚀 Check it out here: arxiv.org/pdf/2408.06724 #AI #DataQuality #Research #SoftwareEngineering #AI

@PAvgeriou This has been well-known to me for a long time; I guess everybody knows this in academia. There are even categories based on name, origin, affiliation, race, etc., which are human factors. The reality is that it is not about the content of the CV but the name on the CV.

We often complain about the #PeerReview system in publications. Things are even worse in evaluating grant proposals: recent evidence shows that panels base their decision mostly on the CV of applicants - the actual proposal makes no difference doi.org/10.1007/s11192…

@shukat_ali I nominate you :)

Call for Nominations for ACM TOSEM's Editor-in-Chief The term of the current Editor-in-Chief (EiC) of the ACM TOSEM is ending on December 31, 2024, and the community is looking for a new EiC. Details here: dl.acm.org/pb-assets/stat…