AI & ML Research 3 Days

Last updated: July 26, 2026, 2:30 PM ET

LLM Agent Development & Optimization

New techniques are emerging to enhance the capabilities of Large Language Model agents. One approach focuses on using OpenAI Agents SDK and Playwright MCP, enabling them to interact with web content. For efficient interaction with specialized models effectively can maximize performance. Addressing a common LLM limitation, a has been developed to improve AI agent memory by prioritizing important information over recent data, inspired by the Ebbinghaus forgetting curve.

Advanced AI & Data Infrastructure

Innovations in data processing and retrieval are crucial for scaling AI applications. Optimizing vector becomes critical when RAM costs rise, exploring trade-offs between on-disk and in-memory Approximate Nearest Neighbor (ANN) indexes like HNSW, SPANN, and Disk ANN. In the realm of tabular data are introduced as foundation models capable of predicting missing spreadsheet columns zero-shot, even outperforming tuned gradient-boosted trees on benchmarks. Furthermore, an can be built and run in the cloud for automating PII classification and extraction from emails using AWS.

AI for Scientific & Engineering Applications

Beyond traditional data tasks, AI is being applied to complex scientific simulations and engineering challenges. A fluid simulator has been developed that generates phenomena like the Kármán vortex street without directly solving fluid equations, utilizing the Lattice Boltzmann Method implemented in C++. In a related area of bioengineering, efforts are underway to, addressing the critical shortage of donor organs by extending their viability. This includes breakthroughs like into pigs, a significant step in organ preservation for transplantation.