Skip to content
Connor KrollsPortfolio

Skills

Every skill here is backed by a project or role on this site.

All skills · 41

  • Python

    Used in Poker AI, Undertone, TEKS Classifier, and UH ML research.

  • Dart

    Used in Undertone.

  • SQL (SQLite)

    Used in Undertone.

  • HTML/CSS

    Used in TSPE.

  • PyTorch

    Used in Poker AI, Undertone, and TEKS Classifier.

  • TensorFlow

    Used in UH ML research.

  • Keras

    Used in UH ML research.

  • scikit-learn

    Used in Poker AI and UH ML research.

  • Hugging Face Transformers

    Used in Undertone and TEKS Classifier.

  • Ollama

    Used in Undertone.

  • NumPy

    Used in Poker AI and Undertone.

  • SciPy

    Used in Poker AI.

  • pandas

    Data wrangling and analysis for research and project experiments.

  • matplotlib

    Used in Poker AI.

  • Deep learning

    A dual-branch CNN-LSTM for glucose forecasting, and Deep CFR in the Poker AI.

  • CNNs

    Used in UH ML research.

  • LSTMs/GRUs

    Used in Poker AI and UH ML research.

  • LLMs and structured output

    Schema-constrained mood tagging with a local LLM in Undertone.

  • NLP and fine-tuning (DistilBERT)

    Used in TEKS Classifier.

  • Generative audio (MusicGen)

    Used in Undertone.

  • Gradient boosting

    Used in Poker AI and UH ML research.

  • Time-series forecasting

    Used in UH ML research.

  • Model calibration

    Temperature-scaled archetype probabilities in the Poker AI.

  • Bayesian inference

    A full posterior over opponent ranges drives the Poker AI's recommender.

  • Empirical Bayes

    Beta-binomial shrinkage of small-sample poker stats.

  • Posterior predictive checks

    Flagging opponents the Poker AI's model can't describe.

  • Experimental design and pre-registration

    Used in UH ML research.

  • CFR and Deep CFR

    Game-theory solvers in the Poker AI, checked against an exact LP.

  • Linear programming

    Sequence-form LPs (HiGHS) for exact equilibria in the Poker AI.

  • Git

    Version control for every project on GitHub.

  • GitHub Actions (CI and scheduled jobs)

    Used in TEKS Classifier.

  • pytest

    Used in Poker AI and TEKS Classifier.

  • REST APIs

    Used in TEKS Classifier.

  • Flutter

    Used in Undertone.

  • Cloud GPU jobs (Kaggle)

    Used in Undertone.

  • Reproducible pipelines

    Seeded, one-command reruns in the Poker AI and an unattended nightly job in the TEKS Classifier.

  • SolidWorks CAD

    Used in Texas Aerial Robotics.

  • Finite element analysis (FEA)

    Used in Texas Aerial Robotics.

  • Tolerance stack-up analysis

    Used in Texas Aerial Robotics.

  • 3D printing

    Used in Texas Aerial Robotics.

  • Mechanical assembly

    Used in Texas Aerial Robotics.