Blog
Blog
Notes on models, ranking, and the slightly uncomfortable question of what a system is actually doing.
Generative AI and language models
What If AI Was Never Meant to Learn From Us
If human data isn't the optimal medium — would we even know? We've spent years feeding models everything humans have ever written, and just assumed that was the...
The Evolution of Reward Hacking and Jailbreak Research in AI
From specification gaming in classical RL to deceptive alignment and jailbreaks in LLMs—a survey of how reward hacking has become a central challenge in AI safety and alignment....
Reasoning in Large Language Models: A Research-Centric Overview
“Can LLMs reason?” is one of those questions that generates more heat than light, mostly because people mean very different things by “reasoning.” This post tries to cut...
Agents
The Web Is Not a Neutral Environment for Agents
Browser agents are getting better fast, but the web is full of things that try to steer behavior. If that already works on humans, why would agents be...
Ranking and recommendation
Modeling Long User Histories for Ads Ranking
How ads ranking systems went from aggregated feature counts to retrieve-and-compress architectures that handle 10,000+ user events under millisecond latency constraints.
Sequential Learning in Recommendation Systems: From Markov Chains to Transformers
A comprehensive guide to sequence-based recommendation techniques and key research papers, tracing the evolution from early statistical methods to modern transformer-based architectures
Contemporary RecSys: Industry-Scale Architectures & Multimodal Systems (2020–2025)
This era represents a pivotal shift toward production-ready, billion-scale architectures that power today’s major platforms. This comprehensive guide covers the essential papers that define contemporary RecSys, from industry-standard...
Deep Learning Era of Ranking & Recommendation Systems: Must-Read Papers (2016–2020)
Explore how deep learning transformed ranking and recommendation systems, from Wide & Deep to BERT4Rec, covering core technologies powering platforms like YouTube, Facebook, and Amazon
Classic Foundational Papers on Ranking & Recommendation Systems
Ranking and recommendation systems power everything from Google Search to Netflix suggestions. While today’s systems use deep learning and large language models (LLMs), their foundations were laid decades...