Showing 30 of 1,445 · Category: Machine Learning
Fast local AI decisions in Rust. A binding for LibLayaX that runs the Laya typed-decision model in-process: no server, no Python. A safe Agent type that is Send + Sync, with no dependencies. Windows, Linux and macOS, x64 and ARM64, CPU or GPU. An open-source alternative to Jev
A statically typed, compiled language for data, ML pipelines and AI serving. Python-like syntax, tensor shapes in the types, typed tables on Arrow, and endpoint functions that become HTTP services.
Alan Vo | Work in progress: identity-risk-workbench-alan-vo
🎬 Intelligent Cinematic Universe & AI Movie Recommendation Engine powered by Latent Semantic NLP, Hybrid SVD Collaborative Filtering, and GSAP/Framer Motion.
Daily download history for every model on the Hugging Face Hub
Evidence-first AI platform for cyber investigation and digital forensics, turning digital traces into traceable, defensible findings.
SYNO: Synthetic Yearning Neural Organism, a brain-inspired AI in pure Python that learns from artificial dopamine and its own bodily needs, with no hard-coded behavior
Vibe Building: a deterministic physics engine, not a language model, decides whether a building design is accepted.
Implementation of a deepfake forensics pipeline combining Fourier spectral analysis, dense optical flow, WavLM audio representations, image-artifact profiling, and vision-language verification with calibrated ensemble scoring.
ML project that predicts EV battery State of Health (SoH, %) using Simple and Multiple Linear Regression in Python (pandas, numpy, matplotlib, seaborn, scikit-learn). Follows the ML life cycle. Uses a synthetic Kaggle dataset.
LOKI — Digital Mentalist is an AI‑powered open source project combining machine learning, psychology, and interactive design to deliver a production‑ready digital mentalism experience.
Azul game AI in Go with a machine-learning brain: alpha-beta search plus an NNUE-style neural network trained on 1.5M self-play positions, +112 Elo over the hand-tuned engine. Go, PyTorch, WebAssembly.
MLOps-Bench evaluates coding agents on repository-grounded production ML engineering changes
Ryan Vo | Work in progress: human-approval-console-ryan-vo
Check image-text SFT data before training: broken records, exact train/eval overlap, and document-disjoint splits.
Heat kills more Indian workers than floods. KisanGuard fights back with 3 ML models, Bedrock Llama 3 advisories in 6 languages, Polly Neural voice alerts on feature phones, A2I human-in-the-loop trust, PPO RL irrigation. 17 AWS services. 1km² ward-level precision.
Clinical cohort data, from raw file to reproducible evidence: data quality, an analysis population with its participant flow, multiple imputation pooled with Rubin's rules, survival models and a Word study dossier. Local-first, open source.
Android malware detection using software-engineering metrics and ML classifiers (course project, ISEC 363)
An intelligent system that senses, learns, understands, and responds to its environment in real time.
Machine Learning End-to-End: Credit Card Fraud Detection - course code for the StackSimplify course
A robust streaming architecture for real-time AI voice agents, featuring dynamic state management and instant interruption recovery.
⚡ NEXA - AI Notice Intelligence System. Automatically parses, prioritizes, scores and extracts actionable deadlines & tasks from campus circulars.
Automated technology intelligence radar and preprint research monitor with keyword scoring and daily Gmail briefings.
A privacy-first Chrome extension for Muslims who want help lowering their gaze online. Uses on-device AI to cover people in images and videos, with customizable masks, blur, and pixelation.
2-bit LLMs for llama.cpp: 85% smaller, beats IQ2_XXS at the same size, no retraining. One-click Windows build.
Tiny byte-level AI on CPU: a zero-dependency C++17 runtime and neural decision models (0.8 ms per 64-byte decision, 9.4 MB RAM). Reads raw bytes, no tokenizer. Specialists for secret scanning (F1 0.797 vs 0.337 for gitleaks on CredData) and PII redaction. Code under Apache-2.0; model weights under their own license.
AI-assisted pre-trade safety layer that turns natural-language trade ideas into validated, risk-aware orders with human confirmation.
V-CARA: Uncertainty-Gated Counterfactual Reinforcement Learning for visual continuous control on an infinite track. Research implementation, experimental results, multi-seed evaluation, and reproducibility resources for visual reinforcement learning.