Serdar Kadioglu

Exactly the goal of our open ended projects! @BrownCSDept Seeing students enjoy it worths every effort you pour into course prep. Here is why I love teaching CS/Optimization:

@dylanjsam @OpenAI @zicokolter @andrew_ilyas @furongh Congrats!!

I defended my PhD thesis! Also, a very (~4 month) late life update, but I've joined @OpenAI to work on safety research and pretraining safer language models! 📈 Thank you to my advisor @zicokolter and my committee: Matt Fredrikson, @andrew_ilyas, and @furongh! 🙏

Forge: Foundational Optimization Representation using Graph Embeddings linkedin.com/posts/serdark_… Paper: arxiv.org/abs/2508.20330

Forge: Foundational Optimization Representations from Graph Embeddings. #Forge is a self-supervised method for embedding optimization problems, just like your favorite text embedding. A single pre-trained model works across tasks, problems, and sizes 🤯 linkedin.com/posts/serdark_…

@charles_irl @modal Applied! Thank you for sharing

The ICLR 2026 deadline is ten days away. But you just found a bug in your evals, so now you need to re-run all your ablations. That's hundreds of experiments, and you need them done ASAP. @modal's got you. Introducing our ICLR 2026 compute grant program.

Balans: Bandit-based Adaptive Large Neighborhood Search ⁦⁦@IJCAIconf⁩ Online, meta-solver for solving mixed-integer programming (MIP) problems. Install: pip install balans Library: github.com/skadio/balans Paper: arxiv.org/abs/2412.14382 linkedin.com/posts/serdark_…

In Montreal for #IJCAI2025? Stop @CSatUSC PhD student Junyang Cai's poster presentation Wed. 10 am. BALANS is an open-source, solver-agnostic meta-solver for #MIPs. Paper: arxiv.org/abs/2412.14382 @BDilkina @serdar_kadioglu @AI4OPT @USCViterbi @IJCAIconf #Optimization #OpenSource

📢 If you're solving real-world problems with AI, we want to hear from you! As Vice Chair of IAAI'26, I invite you to submit your work to Innovative Applications of AI 2026 🚀 🗓️ Deadline: Aug 18, 2025 📍 Conference: Jan 2026, Singapore as part of #AAAI linkedin.com/posts/serdark_…

@lauriewired You just made me remember the summer I spent at AVM, the ActionScript Virtual Machine team at Adobe (really strong engineering team btw) where I studied the AVM specs line by line, and wrote a fuzzer based on grammar guided genetic algorithm —discovered some deep vulnerabilities

I really miss .swf (shockwave flash) as a format. Literally a chain of bezier paths. Scale to any resolution. Oh, and it loads progressively with its own virtual machine (ActionScript Bytecode). No modern filetype comes close. Vector video was a special era.

@siddkaramcheti @ICatGT @GTrobotics @mlatgt Congrats!!

Thrilled to share that I'll be starting as an Assistant Professor at Georgia Tech (@ICatGT / @GTrobotics / @mlatgt) in Fall 2026. My lab will tackle problems in robot learning, multimodal ML, and interaction. I'm recruiting PhD students this next cycle – please apply/reach out!

CORS 2025! June 9–11, 2025, Edmonton, AB

another excellent #NYTechWeek panel with @IBM re: Open Source AI 👏👏👏

@ftm_guney @gorkaydemir @WeidiXie Fantastic news!!

#ICLR decisions are out! we have a paper 🥳 I’m not even gonna try to play it cool 😅 our first ICLR paper, super happy for my student @gorkaydemir and we are extremely lucky to work with @WeidiXie on this project. stay tuned for a very simple online point tracking method with amazing results 🤩

Thank you for the kind invitation @KuisAICenter @ftm_guney —looking forward to it!

@anil_ozturkk richterx.com

parası neyse verip deprem, borsa ve bahis tahmini yapabildiğini zanneden megazeka agaları bir araya getirdiğimiz bir freaksh..forecasting yarışması 🚬

@ftm_guney Super!

yayy I got an outstanding research award at the uni today 🥳 many thanks to our president Metin Hoca and our dean Attila Hoca, not only for the nomination but also for the support 😇

Driven by the complexity and the sheer scale of real-world problems in industry while seeking their abstractions & generalizations. Check out our latest applied research at @Fidelity Ai Center #ai #ml #fidelityassociate linkedin.com/posts/serdark_…

@jmhessel @mervenoyann Sure thing! I wish I could just point us to a tool that already works, but we are getting there.. 🤞

@serdar_kadioglu @mervenoyann Thank you!!

Wanted: LLM that turns language description of problem into convex optimization problem

@mervenoyann @jmhessel Others already mentioned Optverse and Optiguide —which are very cool. And there is an upcoming workshop at AAAI’24 exactly on this topic ai-or.github.io/2024/

@mervenoyann @jmhessel Ner4Opt is avail via ‘pip install ner4opt’ but it only extracts entities without building the complete model. The position paper on Holy Grail 2.0 (CP’23) shows a decomposition guided prompting to generate constraint models —results are very promising arxiv.org/abs/2308.01589