> Thanks for this. #1490 fixed the same deleted-account crash and just merged, so closing this one as a duplicate. Thank you so much for your review.
Leading LLM & agent ecosystems Contributor (Anthropic · Nous Research · SGLang · vLLM · LiteLLM · LangChain) · ships AI agents (archiagents.com) · embedded
🎓 Adaptive AI study agent with POMDP belief state — OPEAA loop, Q-learning + LinUCB bandit policies, SM-2 spaced repetition, concept DAG. Streamlit web app + Chrome extension (MV3). Claude & free HF backends.
NeuroUnfold: Physics-informed deep learning for chirp unfolding in cross-technology wireless sensing. Recovers 406 kHz LoRa beat chirps from 77 kHz aliased BLE RSSI via branch disambiguation, enabling BLE-only respiration sensing.
SWE-bench for your codebase — mine your merged PRs into local, contamination-free coding-agent benchmarks. Adapters: claude-code, aider (Opus 4.7 / GPT-5.5 / Sonnet 4.6 / Gemini 3.1 Pro).
Full-stack agentic web app: LangChain LLM agent + MCP server wrapping a public REST API. Built with Claude Code.
Real-time hand gesture recognition via a Dual-Stream Spatiotemporal Transformer (DSST) on MediaPipe skeletons. Decouples static pose and motion velocity into two Transformer streams with late fusion. Sim-to-Real training: synthetic pretraining (1,200 sequences) + fine-tuning on 35 real clips across 6 gestures. ~557 FPS on a laptop GPU.
High-speed BLE RSSI sampling on nRF54L15 (~77 kHz, 3× prior published rates) for wireless sensing applications. Custom Zephyr firmware that bypasses BLE protocol processing for raw energy detection at the radio level.
End-to-end super-resolution pipeline recovering wideband LoRa channel responses from narrowband BLE RSSI via progressive sub-pixel convolution and physics-informed loss
Build resilient agents.
SGLang is a high-performance serving framework for large language models and multimodal models.
AI Agent to follow NASDAQ trend and generate TQQQ trading signals
- fix: handle null author (deleted/ghost accounts) in GitHub data formatter