Tevfik Kosar

The tide is turning again — There’s Never Been a Better Time to Study Computer Science. An interesting read from The Atlantic.

There is a common misconception in our society, especially among parents and students, that a computer science education is all about programming. No, it is not. To read: tevfikkosar.blogspot.com/2026/05/comput…

Yes, AI Will Replace Programmers. But Only with More Programmers tevfikkosar.blogspot.com/2026/05/yes-ai…

Can you optimize both the performance and energy-efficiency of data transfers at the same time using AI agents? The answer is YES. In our recent paper, we present a dynamic, multi-parameter deep reinforcement learning framework that can achieve this. arxiv.org/pdf/2503.13662

We performed a comprehensive analysis on the energy efficiency of the code generated by 20 popular LLMs.  The study included 878 common programming problems of varying difficulty levels and diverse algorithmic categories. To read the full paper: arxiv.org/abs/2505.20324

In this article, we introduce a novel mechanism to measure network carbon footprints and propose strategies for optimizing the scheduling of network-intensive tasks by shifting them across time and geographic regions based on local carbon intensity.

Our recent article "Towards Carbon-Aware Data Transfers" will be featured in IEEE Internet Computing Magazine. To read the full article: lnkd.in/eC-F2uvk

In this work, we present a dynamic, multiparameter deep reinforcement learning framework that adjusts application-layer transfer settings during data transfers on shared networks. Our method strikes a fine balance between high transfer throughput and low energy utilization.

Our recent paper on “Optimizing Data Transfer Performance and Energy Efficiency with Deep Reinforcement Learning” is now available on arXiv. Any feedback is appreciated. To read the paper: arxiv.org/abs/2503.13662

The paper submission deadline for IEEE SERVICES 2025 conferences -- IEEE CLOUD, EDGE, ICDH, ICWS, QSW, and SSE -- has been extended to March 10th, 2025. Don't miss this opportunity, and submit your work before March 10th! services.conferences.computer.org/2025/

I’m attending Multicore World XII in Christchurch, New Zealand this week. I will be talking about “Embracing AI for Energy-Efficient Data Movement across the Computing Continuum.“ For details: multicore.world/speakers-2025/…

If you are attending AAAS 2025 in Boston this week, don't miss this panel: “Computing Is Eating the World: Will Saving the Planet Destroy It?"   For details: aaas.confex.com/aaas/2025/meet…

I will discuss how to achieve sustainable, scalable, and high-performance computing while mitigating the climate crisis of the Earth. This requires a collaborative effort from academia, industry and government, and a broad adoption of sustainability policies and practices.

I will be speaking at the AAAS’25 panel on “Computing is Eating the World: Will Saving the Planet Destroy It?" this Friday (Feb 14th) at 11:30 am. The panel is organized by CCC Sustainability and Climate Resilience Taskforce. aaas.confex.com/aaas/2025/meet…

A revised and extended version of our work on “How to Evaluate Distributed Coordination Systems? – A Survey and Analysis” is now available on arXiv. Any feedback is appreciated. To read the paper: arxiv.org/pdf/2403.09445

Six IEEE conferences under one roof in a prime location. IEEE World Congress on Services -- hosting IEEE CLOUD, EDGE, ICDH, ICWS, QSW, and SSE -- will take place in Helsinki, Finland from 7-12 July 2025. The deadline for paper submission is March 3rd. lnkd.in/eJqdQmHF

A great venue to publish your work on Sustainable AI and AI for Sustainability. Paper submission deadline is January 15th. We look forward to receiving your submissions! #IEEE CAI 2025 cai.ieee.org/2025/call-for-…

New episode of the Sustainable Computing podcast series: “Carbon-Aware Networking.” You can listen to the podcast at the following links: Spotify: open.spotify.com/episode/6rOfZg… Apple Podcast: podcasts.apple.com/us/podcast/car… Link to our related paper: arxiv.org/pdf/2406.09650 #Sustainability

A mind-provoking poster from NeurIPS - Can an AI Agent Safely Run a Government? And, here ıs the link to the paper: arxiv.org/pdf/2412.00033

We talk about the countless benefits and great potential of AI all the time, but we should also recognize its side effects, such as its significant impact on the environment, and we need to promote responsible and sustainable use of AI. #AI #sustainability