Fast, large scale library for computing rankings and features based on various pairwise and graph algorithms

Experimental Bayesian online ranking system for large-scale quantitative strategy selection.

Developing CausalML uplift model on 6M+ customers to maximize insurance e-sales by through personalized display layout

Agentic room design recommender: retrieval-guided style discovery with generative visualization. Capstone project, 2026

A Retrieval Augmented Generative LLM for Tax Law

GR Training Recipe

In-context learning for Generative Retrieval

Build a recommendation system from scratch

Code for SIGIR 2024 paper: M3oE: Multi-Domain Multi-Task Mixture-of Experts Recommendation Framework

High-Frequency Real-Time Bidding Engine optimized for second-price auctions

Multi-Scenario and Multi-Task Aware Feature Interaction for Recommendation System. TKDD, 2024.

multimodal Click-Through Rate (CTR) prediction task built on the MicroLens-1M dataset (short-video recommendation data). The goal is to predict, for a given user and a candidate video/item, the pro…

使用符号回归在中国股市与加密市场上进行高效因子挖掘。/ By the way, Leverage the novel features and advanced financial mathematics introduced in Uniswap V4 to effectively mitigate just-in-time (JIT) liquidity provision issues.

Large-Scale Trust-Region Methods For Linear Equality Constrained Optimization

[ICML 2024] From Biased Selective Labels to Pseudo-Labels: An Expectation-Maximization Framework for Learning from Biased Decisions

Reproducible MoonEP vs DeepEP v2 dispatch/combine benchmark on 8x B300 (sm_103), with reference results and the environment fixes needed to run both

building-agents-from-scratch

Source code for the dissertation: "Multi-Pass Deep Q-Networks for Reinforcement Learning with Parameterised Action Spaces"

Fast and memory-efficient exact attention

End-to-end training methodology for agent recommendation system

Benchmarking ordinal ranking relevance

This project focuses on improving the personalization capabilities of Large Language Models (LLMs) by advancing the retrieval of user profile information. By using novel retrieval strategies, inclu…