Showing 30 of 1,458 · Category: Machine Learning
From-scratch PyTorch PPO agent that learns to land a simulated 2D rocket: custom NumPy simulator, held-out evaluation with confidence intervals, collapse diagnosis, and a browser replay viewer.
Alan Vo | Work in progress: signal-lens-145-alan-vo
Ryan Vo | Work in progress: signal-pilot-ryan-vo
Alan Vo | Work in progress: model-evaluation-arena-alan-vo
Ryan Vo | Work in progress: nexus-observatory-ryan-vo
Alan Vo | Work in progress: axiom-engine-alan-vo
Alan Vo | Work in progress: axiom-lens-alan-vo
Alan Vo | Work in progress: orbit-hub-144-alan-vo
Ryan Vo | Work in progress: beacon-hub-144-ryan-vo
Ryan Vo | Work in progress: prism-studio-144-ryan-vo
Ryan Vo | Work in progress: sentinel-studio-ryan-vo
Alan Vo | Work in progress: orbit-station-144-alan-vo
Ryan Vo | Work in progress: lumen-foundry-ryan-vo
Zero-day static analysis engine for PE/ELF malware detection using entropy analysis, heuristics, YARA rules, and machine learning.
Machine learning project for predicting the Remaining Useful Life (RUL) of aircraft jet engines using NASA C-MAPSS sensor data, with an interactive Streamlit dashboard for engine health and maintenance insights
Experimental Linux CPU scheduler with fully in-kernel online learning using sched_ext and eBPF.
Filling cloud gaps in Baltic sea temperature with calibrated uncertainty, tested on an unseen basin and the 2018 marine heatwave.
NEXA is an AI-powered adaptive Web Application Firewall that combines Transformer-based request analysis, application context, and session behavior to assess risk, detect suspicious traffic, and explain security decisions.
Browser-based decision tool powered by Typesafe AI Jev using Pollinations API, turning context and structured questions into verdicts, confidence estimates, and score distributions.
Browser-based interview practice tool for Data Science, Machine Learning, and Deep Learning.
Payment fraud detection: realistic simulator with 5 fraud patterns, leak-free streaming features, time-based evaluation, cost-optimal thresholds (82% loss reduction, PR-AUC 0.95) and reason codes.
Research dataset for AI-driven next-24-hour IoMT cybersecurity risk prediction using network, DICOM, and endpoint telemetry features in healthcare environments.
AI & Computer Vision pipeline for rooftop solar capacity estimation using Python, OpenCV, and Scikit-learn.
Fine-tune a chat LLM with LoRA and QLoRA using PEFT, TRL and bitsandbytes. Includes training, inference, adapter merging and a briefing deck.
Open-access course materials on Explainable Artificial Intelligence (XAI), covering foundational concepts, explanation methods, evaluation, and hands-on Python exercises.
Collision-risk monitoring for small satellites: SGP4 screening of the public catalog, an ML risk forecast tested in pre-registered rounds, live space weather and an AI operations assistant.
BHĀRAT-X: Advanced AI & Document Intelligence Foundation for Historical Indic Manuscripts and Cultural Heritage (Phases 1-4)
Credit scorecard (0.60 Gini) with a cost-based cut-off and an approve/decline tool with plain-language reasons