Skip to content
View texasbe2trill's full-sized avatar
🚀
Building systems that turn complex data into actionable insight.
🚀
Building systems that turn complex data into actionable insight.
15 followers · 40 following

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx3Achievement: Public SponsorAchievement: YOLOAchievement: Quickdraw

Achievements

Achievement: Pair ExtraordinaireAchievement: Pull Sharkx3Achievement: Public SponsorAchievement: YOLOAchievement: Quickdraw

Highlights

  • Pro

Block or report texasbe2trill

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.

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
texasbe2trill/README.md

Chris Campbell

typing header

Security engineer by trade. Building and shipping ML, analytics, and AI systems on real datasets.

🎯 What I Do

My day job is security engineering: investigations, telemetry, and incident response. Independently, I design and ship ML and AI projects end-to-end (feature engineering, calibration, validation, deployment) on real datasets with measured outcomes. The projects below reflect that work: real datasets, real metrics, real users.

Currently exploring: conformal prediction intervals, regime-aware recalibration, and LLM-as-judge evaluation harnesses for production model bundles.

205K

row chronological train/test split across 8 NBA targets

R² 0.60 → 0.69

PRA holdout, regular season to playoff bundle

4

shipped open-source projects, 2 with live apps

🚀 Featured Project: Hooplytics

NBA Player Intelligence & Predictive Analytics Platform

Python · scikit-learn · pandas · Streamlit · Typer · joblib

End-to-end ML platform with 60+ pregame-safe features, 8 supervised regressors, market-anchored calibration, regime-aware playoff bundle swap, an 8-page Streamlit app, a Typer CLI, and ReportLab scouting PDFs.

Held-out R² (chronological split, n_test = 41,024):

Target R² MAE Target R² MAE
PRA 0.599 6.06 Rebounds 0.481 1.89
Fantasy 0.556 7.82 Turnovers 0.330 0.87
Points 0.534 4.60 3PM 0.305 0.90
Assists 0.512 1.38 STL+BLK 0.195 0.95




📦 Other Public Work

📚 KoNotes

Python · Streamlit · LLMs · NLP

Local-first AI-assisted knowledge analytics. Converts Kobo and Kindle annotations into structured, queryable insight with explainable, rule-based recommendations.

🚨 AlertSage

Python · Jupyter · scikit-learn

Reference workflow for security alert classification: TF-IDF and lexical features, calibrated thresholds, and a structured evaluation harness for repeatable model comparison.

🍎 macos-trust

Python · CLI

Context-aware macOS trust assessment. Fast evaluation of apps, launch items, and system controls with low false-positive design.

🛠️ Stack
ML / Modeling scikit-learn, classification, regression, anomaly detection, calibration, time-aware validation, residual diagnostics, threshold tuning
Data Science Python, SQL, pandas, NumPy, statistical reasoning, EDA, reproducible Jupyter workflows
Applied AI LLM summarization, structured extraction, text classification, AI-assisted triage, RAG, embeddings
Domain security telemetry, alert triage, signal engineering on auth / network / endpoint, false-positive reduction
Delivery Streamlit dashboards, Typer CLIs, joblib model artifacts, ReportLab reports, REST APIs

📊 GitHub


Pinned Loading

  1. hooplytics hooplytics Public

    Hooplytics turns NBA box-score data into player intelligence through machine learning, interactive analytics, and visual workflows for exploring trends, projections, and performance signals.

    Jupyter Notebook 3

  2. KoNotes KoNotes Public

    Interactive reading dashboard and CLI for Kobo and Kindle — library stats, AI-powered insights, smart recommendations, book cover art, reading activity, annotation exports, and Bluesky sharing.

    Python 4 2

  3. AlertSage AlertSage Public

    An NLP system for classifying cybersecurity incident descriptions into meaningful event types. Designed to mirror early SOC triage, it transforms unstructured analyst text into structured labels us…

    Jupyter Notebook 5 2

  4. macos-trust macos-trust Public

    Intelligent macOS security scanner that identifies unsigned apps, Gatekeeper violations, and suspicious persistence mechanisms with context-aware risk assessment. No false positive fatigue.

    Python 2