Decoding the Super Mario 64 courtyard plaque ("L is real 2401"): font identification, forward model of the 1996 texture, line-by-line decoding, report and full reproduction code.
Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Transparent proxy server that works as a poor man's VPN. Forwards over ssh. Doesn't require admin. Works with Linux and MacOS. Supports DNS tunneling.
A straightforward method for training your LLM, from downloading data to generating text.
๐ซ Ngrok FRP Alternative โข โก Fast โข ๐ชถ Lightweight โข 0๏ธโฃ Dependency โข ๐ Pluggable โข ๐ TLS interception โข ๐ DNS-over-HTTPS โข ๐ฅ Poor Man's VPN โข โช Reverse & โฉ Forward โข ๐ฎ๐ฟ "Proxy Server" framework โข ๐ "Web Server" framework โข โต โถ โท โ "PubSub" framework โข ๐ท "Work" acceptor & executor framework
A comfortable and well-configurable graphical Frontend for incremental backups, with a command-line version also available. Modified files are transferred, while unchanged files are linked to the new folder using rsync's hard link feature, saving storage space. Restoring is straightforward via file manager, command line or Back In Time itself.
Designed for better message forwarding in Telegram.
Capture every activation and gradient of any PyTorch model โ forward and backward โ with automatic graph visualization, rich metadata, and live interventions. Works on any architecture, including dynamic and recurrent ones.
๐ OSINT Maltego Transforms for investigating Telegram channels, groups, and users, including deanonymization via stickers, forwarded messages, similar channels, deleted posts, and more.
InfiniSplat v1 & v2: Feed-Forward 3D Gaussian Reconstruction
AgentJev-0.6B - a fast 'System One' decision model for AI Agents: feed it any unstructured state (diffs, traces, logs) and structured questions, get calibrated probability distributions back in one ~50ms forward pass. Zero output-token decoding.
A curated collection of resources, tools, practices, case studies, and opportunities for Forward Deployed Engineers.
We propose LeapTalk, a novel framework that achieves stable and real-time talking-head generation with a single forward step, scaling to arbitrarily long videos.
๐๐ A multimodal System One decision model that gives calibrated answers to typed questions about screens, photos, video and text in one forward pass.
Forward videos from IM to transcript, Markdown, and your knowledge base โ with raw and skill-updated outputs.
JevK5: open-weight alternative to TypeSafe Jev. Typed decisions with probabilities in one forward pass; Apache-2.0 weights and code.
NVIDIA Sol-Attn for ComfyUI / Triton kernel on SM89 - SM121, with zero-copy MiniMax H3 nodes: memory-efficient attention, scheduled tau with graph preview, and feed-forward chunking. Measured 1.14โ1.44ร vs SageAttention and โ37% MLP peak VRAM on H3
Open-source FDE project workbench for research, AI opportunities, delivery and team collaboration.
Jev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.
Open-source alternative to TypeSafe's Jev: a System One style model layer that gives typed, calibrated decisions from any open-weights LLM in one forward pass (HF + vLLM), with honest benchmarks
Metal-native differentiable 3D Gaussian Splatting for Apple Silicon: fused forward+backward rasterizer on PyTorch MPS, no CUDA anywhere.
Xray desktop client that tunnels your whole system: split routing (Iran/Russia/China), per-username multi-exit SOCKS, Wi-Fi hotspot sharing, port forwarding, anti-filter tuning (fragmentation, ECH, REALITY PQ), low-usage mode.Windows, Linux & macOS..
[ECCV'26] Map-Det3D: Metric Feed-Forward 3D Reconstruction Prior for Multi-view 3D Object Detection from Streaming Inputs
Local Responses API proxy for the ChatGPT Excel backend, with client-side tool forwarding.
Source-backed, production-first FDE interview fieldbook with cases, scorecards, and current role research.
```text Interactive A-share strategy research with staged entries, walk-forward validation, trade-level explanations and offline reports. Built with Python, React and FastAPI.
Dev (dev-0.4b): a 0.4B bidirectional decision model on ModernBERT-large for routing, yes/no verification, and 1-5 rating. Non-generative single forward pass with calibrated confidence.
Open replica of TypeSafe's Jev: typed calibrated decisions in one forward pass, on Gemma 4 E2B / Gemma 3 270M (Modal)
MoJev: typed, calibrated decisions in one forward pass.
Valida ideas de trading como lo harรญa un fondo: 6 agentes de IA, el mรฉtodo TIS y un motor de 5 fases. Trae guรญa para quien no programa y 7 estrategias de ejemplo, las 7 descartadas con motivo escrito.
Production & research-grade Post-Quantum KEM Double Ratchet protocol (FIPS 203 ML-KEM-768 + FIPS 204 ML-DSA-65). Resilient against Harvest-Now-Decrypt-Later FTQC attacks
A tiny jev-like model that answers Choice, Score and Noul questions in one forward pass and returns calibrated probabilities. MLX or PyTorch, fully offline, System One compatible.
100+ AI, Gen AI, LLM, Agentic AI, Forward Deployed Engineer (FDE), AI Systems Architect, Applied AI, LLMOps, AI Platform engineering interview questions across 10 topics - short answer, detail, worked example, and interview tips for each.
MCP server + NinjaTrader 8 AddOn that gives an AI coding assistant eyes and a build loop for NinjaScript: compile with structured errors, chart and Print reads, headless backtests, walk-forward, lint, trade match vs research, screenshots. Opt-in Sim-only order and strategy control, plus Sim desk crash-recovery tools. Read-only by default.
[ACM MM 2026] PURE: Refining Feed-forward 3D Reconstruction with Unlabeled Post-training