champion19007 champion19007

My work sits at the intersection of LLMs, Computer Vision, and Reinforcement Learning — I design agents that reason, plan, see, and act with autonomy.

I have worked across the stack: transformer models and RAG pipelines, ROS2 robotics systems, CV model training, and hybrid RL + search algorithms. I care about engineering that is clean, scalable, and grounded in real research.

Currently exploring — how search, reinforcement learning, and transformers combine into something greater than their parts.

_{resting at the bench}

_{sealed vessels — the work worth opening first}

♟️ Hybrid Minimax–MCTS Chess Engine A patent-inspired hybrid engine fusing Minimax, MCTS, and deep evaluation networks in C++.	🤖 RL Chess Engine — Deep RL + Search Reinforcement learning chess engine pairing policy/value networks with Monte Carlo Tree Search.
🐾 Real-Time Animal Detection (CCTV) YOLO + OpenCV pipeline detecting animals live from CCTV streams — built for safety monitoring and rapid alerting.	📈 Advanced Stock Forecaster Hybrid CNN + BiLSTM + GRU model for time-series forecasting, capturing local patterns and long-range dependencies.

⟡ Languages

⟡ Frameworks

⟡ Tools & Platforms

⚔️ What I Build

LLM Agent Systems — planning, tool use, memory
RAG & Retrieval Pipelines — vector search, rerankers
Computer Vision Models — CNNs, ViTs, YOLO
Reinforcement Learning + Search — MCTS, Minimax
Backend Architectures — FastAPI, PostgreSQL, Docker
Hybrid Research Engines — Search + RL + Transformers

✨ What I Can Help With

Agent design — knowledge graphs, tool routing, memory
Vision systems — detection, segmentation, transformers
Reward engineering — game-AI, hybrid MCTS models
Model serving — scalable inference backends
PyTorch optimization — speed, memory, ONNX export
MLOps — tracking, CI/CD, reproducible pipelines

_{⟡ RELICS OF THE HUNT ⟡}

_{⟡ TRAILS WALKED ⟡}

◆ Full metrics — click to expand

_{notes → documentation → experiments → deployment}

◆ My research & documentation workflow

📚 Notes & Knowledge Base

Obsidian — main AI/ML knowledge vault
Notion — structured docs + planning
Excalidraw — ML architecture sketches
Draw.io — system & pipeline diagrams

🎥 Learning Channels

Andrej Karpathy — transformers, LLM systems
Yannic Kilcher — paper deep dives
Two Minute Papers — research summaries
DeepMind — research talks

📄 Reference Documentation

🧪 Experimentation & Tracking

Weights & Biases — experiment tracking
TensorBoard — training visualization
Jupyter — exploration notebooks

🔧 Useful Tools

Papers With Code · arXiv
Ollama · LM Studio

_{⬦ Hallownest awaits ⬦}

Navigation Menu

champion19007 champion19007

Achievements

Achievements

Block or report champion19007

♟️ Hybrid Minimax–MCTS Chess Engine

🤖 RL Chess Engine — Deep RL + Search

🐾 Real-Time Animal Detection (CCTV)

📈 Advanced Stock Forecaster