Pinned Loading

1. rag-from-scratch rag-from-scratch Public

   RAG pipeline from scratch — FAISS + Sentence-Transformers + Mistral 7B via Ollama. No LangChain. FastAPI backend with chunking benchmarks, latency profiling, and LLM abstraction layer.

   Python

2. autonomous-ai-agent autonomous-ai-agent Public

   Autonomous ReAct agent built from scratch — no LangChain. Features circuit breaker, dual database (ChromaDB + SQLite), Prometheus/Grafana monitoring, and a provider-agnostic LLM layer (Ollama/OpenA…

   Python

3. production-rag production-rag Public

   Production RAG system with LangChain, Qdrant & Docker. Multi-user isolation, 3 retrieval strategies (Similarity/MMR/HyDE), reranking, query rewriting, async ingestion worker.

   Python 1

4. fine-tuned-with-lora fine-tuned-with-lora Public

   LoRA fine-tuning of TinyLlama on coding instructions · MLflow registry · SSE streaming FastAPI · Blue-green Docker deployment · 37% perplexity improvement

   Python 1

5. multi-agent-system multi-agent-system Public

   Production multi-agent system: LangGraph supervisor graph, MCP tool server, ChromaDB memory, SQLite observability, FastAPI — runs free on local Ollama.

   Python

6. advanced-multi-object-tracking advanced-multi-object-tracking Public

   Advanced multi-object tracking pipeline — YOLOv8m + StrongSORT with OSNet appearance embeddings on MOT17. HOTA 41.6 | MOTA 38.1 | IDF1 50.8. Benchmarked against ByteTrack baseline with TrackEval.

   Python 1