Showing 30 of 1,450 · Category: Machine Learning
Ryan Vo | Work in progress: image-quality-observatory-ryan-vo
An NLP-powered web app that detects spam in emails and SMS messages using a Multinomial Naive Bayes classifier with TF-IDF features.
Predictive maintenance for small factories. MachinaSense detects bearing and motor faults from vibration and temperature data using unsupervised AI trained only on normal data. Features English + Hindi alerts, voice/beep warnings, live graphs, CSV diagnostics and an AI chatbot. Edge-ready, ~₹1,800 per motor.
StayGuard AI is an enterprise-grade, end-to-end MLOps pipeline designed to predict hotel reservation cancellations with high precision. Features schema validation via Pandera, automated experiment tracking with MLflow, containerized Flask serving, and automated CI/CD via Jenkins & Docker on GCP.
Text-to-SQL with a 852k-parameter transformer trained from scratch, compared with LoRA and a 120B API model on identical questions.
Structured Python practice covering fundamentals, data visualization, web scraping, PySpark, and machine learning.
Lattice Representation Hypothesis: code and results for 'From Plato to Aristotle: Toward a General Theory of Representation Hypotheses'
Alan Vo | Work in progress: cobalt-studio-alan-vo
Ryan Vo | Work in progress: lumen-workbench-ryan-vo
Alan Vo | Work in progress: orbit-workbench-144-alan-vo
Alan Vo | Work in progress: prism-lab-alan-vo
Ryan Vo | Work in progress: orbit-engine-145-ryan-vo
Re-derivation of oxygen-vacancy formation-energy results in non-magnetic oxides with structure-only and descriptor-augmented graph networks: stored records, predictions, specs and scripts; no GPU or retraining needed.
Verified JAX/Equinox ports of Hugging Face LLMs and embedding models: numerically matched against PyTorch, forward and backward, and saved back to Hugging Face formats.
Pre-load safety scanner for downloaded ML models: inspects artifacts for code-execution payloads, unsafe serialization (pickle), and backdoor signatures before from_pretrained() runs. Static pickle opcode analysis + format rating + repo/metadata audit → SAFE/SUSPECT/UNSAFE verdict with SARIF output.
Ryan Vo | Work in progress: beacon-workbench-ryan-vo
Ryan Vo | Work in progress: atlas-lens-ryan-vo
Ryan Vo | Work in progress: orbit-workbench-ryan-vo
Multi-agent match intelligence built on synthetic football event data (no real Premier League data used). Built for the Inside the Game: Developer Hackathon.
Ryan Vo | Work in progress: nexus-console-ryan-vo
PyTorch implementation for training JEV-style calibrated decision models
A research harness built on Pi: reproducible ML experiments, traceable evidence, bibliographic checks and scientific writing.
AI resume vs JD analyzer - TF-IDF, skill ontology, match score & project suggestions
Ryan Vo | Work in progress: cobalt-console-ryan-vo
Alan Vo | Work in progress: signal-forge-144-alan-vo
Alan Vo | Work in progress: orbit-lab-alan-vo
Ryan Vo | Work in progress: cobalt-lab-144-ryan-vo
Alan Vo | Work in progress: nexus-hub-alan-vo
Twin-Pulse: Cardiometabolic Digital Twin Workstation — Submission for the Happiest Health Digital Twin Challenge 2026