LLM Powered Apps Tech Stack
LLMs have emerged as an influential and versatile tool for software development, as hinted by Andrej Karpathy in his article Software 2.0 This post illuminates the tools and options for each stage of the pipeline, heads up there is a lot of them. I will try to keep it as simple as possible.
Handling Data
| Data Pieplines | Vector Databases | LLM Cache |
|---|---|---|
| Databricks | Weaviate | Redis |
| Airflow | Chroma and Faiss | SQLite |
| Unstructured | Pinecone | GPTCache |
| pgvector | LMQL |
Constructing app logic
| LLMs & Embedding | Orchestration | AGI |
|---|---|---|
| OpenAI | LangChain | Auto GPT |
| Anthropic | LlamaIndex | BabyAGI |
| Hugging Face | Semantic Kernel | AgentGPT |
| Open LLMs | HayStack |
Execution and Inference
| LLMops | App Hosting | APIs & Plugins |
|---|---|---|
| Agenta | Vercel | ChatGPT Plugins |
| MLFlow | Streamlit | Zapier |
| Comet ML | Gradio | Botpress |
| W & B | Modal | Serp |
Extra Resources
Conclusion
I hope this post helps you navigate the LLM landscape and build your next LLM powered app.