Hui-Kang Tong
Experience
Quora
Nov 2021 - Present
Machine Learning Engineer, San Francisco Bay Area
- Main maintainer of AI tooling for the company monorepo
- Established the internal agent skills system and feedback loops
-
Created CLIs and endpoints for AI tools to access almost all internal information
- Designs, implements and A/B tests changes for Quora feed and digest recommendations
- Built tools like funnel analysis to investigate any issues with retrieval and ranking algorithms
-
Runs the product interview for content recommendation systems and AI engineering
- Shipped the generation of questions about recent events
-
Wrote and evaluated prompts to label topics for content, and the pipeline to create new topics
- Languages: Python, agent skills, Trino/PrestoSQL, PyTorch
- Technical domains: AI engineering, LLMs, AI tooling, recommendation systems, A/B testing
Projects
NVIDIA Nemotron Model Reasoning Challenge
Won the Open Progress Prize (publication, demo) (April 2026)
ARC Prize
Ranked 4th at milestone 1 of the ongoing ARC-AGI-3
Worked on ARC-AGI-2 in 2025, crafted solutions (writeup)
AI Mathematical Olympiad
Published notebooks and annotation corpus for AIMO 3 in 2026, commended by the hosts
Previously participated in AIMO 1 (code, discussions) and AIMO 2 (code, writeup)
Konwinski Prize 2025
Published the widely used baseline notebook. Ranked 7th on the public leaderboard
Meta AI Hacker Cup
Competed in Meta’s AI competitive programming challenge in 2024 (3rd place) and 2025
LeetCode - peak rating 2992 (March 2022)
Codeforces - peak rating 2235 (July 2022)
Writing
Predictions - 2026, interaction models and multichannel prediction, supercommunity
Product - instruction delivery, principles for AI products, prompting projects
Drawings - animated flash attention, decoder layer, X algorithm
AI Coding - mindset changes, what AI needs, how to improve
Education
Singapore University of Technology and Design
May 2017 - Aug 2021
Engineering Systems and Design, Minor in Computer Science, GPA 4.61/5.00