Skip to content
View Darrenus's full-sized avatar
🎯
Focusing
🎯
Focusing
6 followers · 16 following

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx2Achievement: YOLO

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx2Achievement: YOLO

Block or report Darrenus

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Darrenus/README.md

Rong He · 贺融

Human-centered AI-nerd

Hi, I'm Rong 👋 I build AI agents and backend systems during my school work and intern. Currently studying MTech in software engineering at NUS in Singapore, after B.S. in computer science at KAIST.

Nowadays, I'm using AI to explore unfamiliar fields, connect ideas, and turn real-world problems into practical solutions. I’m curious about what one person can learn, create, and accomplish with AI.

These days, I'm building a coding agent from scratch, exploring reproducible quantitative research with TickWitness, and developing Lumen Next an auto-research agent. I enjoy learning about new domains, identifying meaningful problems, and using AI to turn ideas into something useful.

Languages
Python SQL TypeScript Java C++

Agents & retrieval
Planner–Executor LangChain MCP Hybrid retrieval KV-cache reuse

Backend & data
FastAPI Async queues MySQL MongoDB Redis MinIO

Evaluation
A/A baselines NDCG Ablations

Computer vision
YOLOv8 Graph neural networks GPU inference

 Rong's GitHub statistics: public contributions over the past 12 months

Featured work

coding-agent-from-scratch — Independent open-source agent with repository maps, sandboxing and controlled evaluation TickWitness — Reproducible quantitative research with replay experiments and independently verifiable evidence Lumen — AI-assisted mathematical research with parallel agents, literature search and computation

AI Workload

AI workload: Codex and Claude Code token shares over the same 28 days, daily activity heatmaps, cumulative tokens, weekly averages and call frequency

Local Codex & Claude Code activity. Tokens include cached input; weekly averages use the last 28 days. Web chats and other devices are outside this snapshot. Counting method ↗

Research

Agents & LLM systems

Learning theory & mathematics

All works above are preprints. Research notes & full records ↗

Website  ·  Résumé  ·  LinkedIn  ·  Email

Pinned Loading

  1. coding-agent-from-scratch coding-agent-from-scratch Public

    从零手写的终端 coding agent（无框架）：ReAct 循环、SEARCH/REPLACE 匹配策略链、tree-sitter + PageRank 仓库检索、Seatbelt 沙箱。每个设计决策都由带 A/A 噪声校准的对照实验验证。

    Python

  2. tickwitness tickwitness Public

    Find reproducible counterexamples to quantitative research claims under explicit data and execution assumptions.

    Python

  3. ziyangliu-666/lumen-next ziyangliu-666/lumen-next Public

    An AI that does research in mathematics on its own. It reads papers, runs code in a sandbox, and you can watch and steer it while it works.

    TypeScript 3