Skip to content
View yz5262-afk's full-sized avatar

Block or report yz5262-afk

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. Markdown is supported. 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
yz5262-afk/README.md

Jonathan Zhang

Quantitative research · Optimization · Machine learning

I build research tools that make models inspectable: from Monte Carlo risk surfaces and integer dynamic programs to causal attention, game-tree search and dependence-aware statistics.

Selected projects

Project Focus
Snowball Risk Studio Monte Carlo autocallable valuation, common-random-number Greeks, barrier analytics and a Streamlit risk dashboard.
Execution DP Lab Integer dynamic programming for trade scheduling with memory-dependent fills, dominance pruning and exact small-instance checks.
Keeling Curve Lab Reproducible NOAA CO2 trend and seasonality analysis with calendar-aware HAC covariance, sensitivity studies and an independent growth benchmark.
MVO Research Lab Clustering-driven equity preselection, mean-variance allocation and downside-oriented portfolio research, with a standalone synthetic demo.
Transformer Language Lab PyTorch attention and causal Transformer experiments with reusable modules, tokenization comparisons and prefix-invariance checks.
Music Sequence Lab RNN/LSTM symbolic music experiments with song-level splits, repetition-aware sampling and MIDI export.
Statistical Learning Lab Ridge and kernel-ridge experiments with regularization selection, fold-local preprocessing and reproducible synthetic comparisons.
Akuna Market Making Lab Competition-derived binary-options market making with parameter estimation, Monte Carlo probabilities, inventory-aware quotes and FOK risk gates.
Hexapawn Search Lab Java Hexapawn with depth-limited expected-outcome search, seeded AI matches and independent checks across reachable game states.

Research approach

  • Make assumptions visible. Model boundaries and source context accompany the code.
  • Check what can be checked. Exhaustive small cases, causal-prefix tests and numeric constraints support the implementations.
  • Keep examples reproducible. Public-data and synthetic demos can be explored without employer files.

The repositories include independent research, coursework-derived implementations and contributions to collaborative workflows. Each project describes its provenance, what was verified and what still depends on external data or longer training runs.

Tools: Python · Java · NumPy · SciPy · pandas · PyTorch · scikit-learn · Streamlit

Pinned Loading

  1. snowball-risk-studio snowball-risk-studio Public

    Monte Carlo autocallable valuation, common-random-number Greeks, barrier analytics and a Streamlit risk dashboard.

    Python

  2. execution-dp-lab execution-dp-lab Public

    Integer dynamic programming for trade scheduling with memory-dependent fills, dominance pruning and exact small-instance checks.

    Jupyter Notebook

  3. mvo-research-lab mvo-research-lab Public

    Clustering-driven equity preselection, mean-variance allocation and downside-oriented portfolio research, with a standalone synthetic demo.

    Jupyter Notebook

  4. keeling-curve-lab keeling-curve-lab Public

    Reproducible NOAA CO2 trend and seasonality analysis with calendar-aware HAC covariance, sensitivity studies and an independent growth benchmark.

    Python

  5. transformer-language-lab transformer-language-lab Public

    PyTorch attention and causal Transformer experiments with reusable modules, tokenization comparisons and prefix-invariance checks.

    Jupyter Notebook

  6. music-sequence-lab music-sequence-lab Public

    RNN/LSTM symbolic music experiments with song-level splits, repetition-aware sampling and MIDI export.

    Jupyter Notebook