Chaoyu Wang
王超宇
AI Researcher & Full Stack Developer Co-Founder at MAL.LAB
About
Hi, I'm Chaoyu 👋, Founder at MAL.LAB and independent ML researcher — interested in building AI systems that are more capable, more reliable, and genuinely beneficial to humanity.
I received an M.S. from 
Northwestern University and a B.S. from 
UC San Diego, both in Applied Mathematics — two places I am deeply grateful for. I was fortunate to be mentored by Prof. Zhaoran Wang at Northwestern and Prof. Ioana Dumitriu at UCSD, both of whom shaped how I think about research. My previous work spans LLM fine-tuning and alignment, retrieval-augmented generation, and synthetic data construction.
You can find more about my background in my CV.
🔬 Interests: LLM Alignment & Safety, Reward Modeling, Preference Optimization, Evaluation & Interpretability.
More than a position, I am looking for the right fit — a lab that takes its time with ideas, a collaborator who wants to build something meaningful over the long run, or a mentor genuinely invested in helping someone learn to think independently. I believe this kind of match has to go both ways.
I am available to work fully onsite for six months or more, and I take that commitment seriously — good research takes time, and I am not looking to pass through.
If any of this resonates, I would love to talk: email · calendly
Latest News
Started building my PhD application portfolio at UC Berkeley
Started building my PhD application portfolio at UC Berkeley
📚 Now based in the Bay Area, self-studying daily at UC Berkeley libraries and engaging with the research community. Targeting CS PhD (Fall 2027) with a focus on LLM.
Launchpad S1 successfully held in Shanghai
Launchpad S1 successfully held in Shanghai
🚀 Co-organized Launchpad S1 with MAL.LAB — a product launch and go-to-market event in Minhang, Shanghai. Brought together founders and builders, sponsored by DeepTech. From idea to execution in two months.
Joined GuruGame HK as ML Engineer Intern
Joined GuruGame HK as ML Engineer Intern
🎮 Started building an ad campaign agent with SFT fine-tuning, tool-calling, and RAG pipeline. Released AdCampaignAgent-SFT dataset on HuggingFace.
Graduated from Northwestern University
Graduated from Northwestern University
🎓 Completed M.S. in Engineering Science & Applied Mathematics at Northwestern University.
Joined Huatai Securities as Software Dev Intern
Joined Huatai Securities as Software Dev Intern
📈 Built a RAG-based Q&A system over 2,000+ financial documents and applied GRPO alignment on DeepSeek-R1-7B, reducing hallucination rate from 8.0% to 1.2%.
Check out my latest work
AdCampaignAgent-SFT
AdCampaignAgent-SFT
Rule-based synthetic SFT dataset for mobile game UA agents, featuring tool-calling, multi-turn reasoning chains, and ROAS/Retention safety baselines. Fine-tuned Qwen3-0.6B with LoRA, achieving 86%+ end-to-end task completion.
FinReas-R1
FinReas-R1
Reasoning Reward Model trained via GRPO on synthetic Chinese customer service preference data. Generates evaluation rationale before outputting preference labels, reducing reward hacking vs. scalar RM. Built on DeepSeek-R1-Distill with veRL + vLLM rollout pipeline.
Launchpad S1 @ Shanghai
Launchpad S1 @ Shanghai
As Founder of MAL.LAB, I organized Launchpad S1 a 72-hour GTM sprint held in Shanghai Minhang. The event brought together 200+ participants, 30+ teams, and 20+ ecosystem sponsors, resulting in 30+ live product launches, ~200 pieces of launch collateral, and 50,000+ impressions on Xiaohongshu. Guests included investors, corporate partners, and media.
Taste For Agents Not Human
Taste For Agents Not Human
Taste is the primary information source for AI agents. It is a living knowledge base of tools, workflows, and expert insights — discoverable through a simple CLI, usable in any domain.
PromptGen - TextToImg Pipeline
PromptGen - TextToImg Pipeline
AI-powered prompt generation and image production system with template-driven workflows, multi-provider orchestration, and multilingual image stitching.
Monitor System For User Acquisition
Monitor System For User Acquisition
Internal full-stack monitoring system for Google Ads operations, AppsFlyer cohort analytics, and evaluation-driven optimization workflows.
