
Adam Ru Kun Dong
AI research and quantitative research
I am Adam Dong, an MSc student at TU Delft. I study how trading agents handle risk and build open-source tools for AI agents.
Recent work
- Risk-sensitive RL for trading
Risk-sensitive RL for trading
How quadratic risk penalties change a market maker and an execution agent that learn together on Amazon order-book data.
- Execution: 59% less remaining quantity, completion from 95.3% to 99.9%
- Market maker: 48% less inventory at a moderate penalty, then it stops trading
Open the write-up
- ros.ai
ros.ai
An open-source workspace where AI agents from any vendor work on one shared canvas, with its own fine-tuned local model.
- NVIDIA Nemotron 3 Nano 4B fine-tuned with LoRA on about 3,000 agent turns
- 68 of 95 held-out tasks passed, against 33 for the base model
Open the write-up
- AURORA
AURORA
A pipeline that grounds a draft in approved ABN AMRO sources before it is written, built with AISO.
- Retrieval, citation checks and KPI-based evaluation
- A REST API, eight MCP tools and a Streamlit interface
Open the write-up
Education
- MSc Data Science and AI Technology
MSc Data Science and Artificial Intelligence Technology
Delft · September 2026 to September 2028 (expected)
Themes: Advanced Machine Learning and Probabilistic Decision Making. Theme curriculum:
- Advanced Machine Learning: Elements of Statistical Learning, Alternative Learning Strategies, Generative Modeling
- Probabilistic Decision Making: Probabilistic Models and Inference, Sequential Decision Making, Deep Reinforcement Learning
- BSc Computer Science
BSc Computer Science
Amsterdam · September 2022 to September 2026
Coursework included:
- Scientific Data Analysis
- Machine Learning
- Algorithms and Complexity
- Thesis on risk-sensitive multi-agent RL for trading
Leadership
- Head of Research
Head of Research
Amsterdam · May 2026 to now
I lead the technical AI and ML events, and have mentored student projects and a quantitative-finance track.
- Head of AI and ML Research
Head of AI and ML Research
Amsterdam · July 2026 to now
I lead the AI and ML research agenda for quantitative and agentic trading systems.
Skills
Python
C++
JAX
PyTorch
Hugging Face
Docker
Recent work
- Risk-sensitive RL for trading
Risk-sensitive RL for trading
How quadratic risk penalties change a market maker and an execution agent that learn together on Amazon order-book data.
- Execution: 59% less remaining quantity, completion from 95.3% to 99.9%
- Market maker: 48% less inventory at a moderate penalty, then it stops trading
Open the write-up
- ros.ai
ros.ai
An open-source workspace where AI agents from any vendor work on one shared canvas, with its own fine-tuned local model.
- NVIDIA Nemotron 3 Nano 4B fine-tuned with LoRA on about 3,000 agent turns
- 68 of 95 held-out tasks passed, against 33 for the base model
Open the write-up
- AURORA
AURORA
A pipeline that grounds a draft in approved ABN AMRO sources before it is written, built with AISO.
- Retrieval, citation checks and KPI-based evaluation
- A REST API, eight MCP tools and a Streamlit interface
Open the write-up
Education
- MSc Data Science and AI Technology
MSc Data Science and Artificial Intelligence Technology
Delft · September 2026 to September 2028 (expected)
Themes: Advanced Machine Learning and Probabilistic Decision Making. Theme curriculum:
- Advanced Machine Learning: Elements of Statistical Learning, Alternative Learning Strategies, Generative Modeling
- Probabilistic Decision Making: Probabilistic Models and Inference, Sequential Decision Making, Deep Reinforcement Learning
- BSc Computer Science
BSc Computer Science
Amsterdam · September 2022 to September 2026
Coursework included:
- Scientific Data Analysis
- Machine Learning
- Algorithms and Complexity
- Thesis on risk-sensitive multi-agent RL for trading
Leadership
- Head of Research
Head of Research
Amsterdam · May 2026 to now
I lead the technical AI and ML events, and have mentored student projects and a quantitative-finance track.
- Head of AI and ML Research
Head of AI and ML Research
Amsterdam · July 2026 to now
I lead the AI and ML research agenda for quantitative and agentic trading systems.
Skills
Python
C++
JAX
PyTorch
Hugging Face
Docker
Selected projects
See all 10 write-ups
ros.ai
2026An open-source workspace that holds a whole project on one canvas, where AI agents from any vendor work from the same blocks, with its own fine-tuned local model.

Risk-sensitive RL for trading
2026My bachelor thesis: how quadratic risk penalties change a market maker and an execution agent that learn together on Amazon limit-order-book data.

AURORA
2026A proof of concept for ABN AMRO: a pipeline that turns a rough writing request into a draft with retrieved sources, quality checks and a trace of how it was made.
Latest blog posts
Read all 4 posts- From Vague Prompts to Grounded Content
AURORA, part 2: a seven-stage grounding pipeline.
24 May 2026 - A Controlled Study of Risk Aversion
Thesis, part 2: quadratic risk penalties.
24 May 2026 - Making Trading Agents Care About Risk
Thesis, part 1: three risk-sensitive objectives.
28 Apr 2026 - An AI Editorial Co-Pilot for ABN AMRO Insights
AURORA, part 1: why retrieval matters most.
20 Apr 2026

Focus
I am motivated by problems of uncertainty, information and incentives, and I aim for a career in quantitative research, trading or risk management, where rigorous modeling and real-world impact meet.