our tech & tools

here's the toolkit that powers our ai-driven workflows—from data ingestion and realtime streams to seamless llm ops—complete with links to get you started in seconds.

primary technologies

• python (python.org) serves as our backbone for scripting, data processing, and glue code across services
• apache kafka (kafka.apache.org) powers our high-throughput event streams and decoupled pipelines
• postgresql (postgresql.org/docs) provides reliable relational storage and time-series snapshots for analytics
• supabase (supabase.com/docs) offers instant postgres APIs and real‑time subscriptions for our frontend
• next.js (nextjs.org) delivers our react‑based ui with server‑side rendering and static exports

ai toolkit

• langchain (langchain.com) orchestrates llm calls into pipelines, agents, and chains
• openai (openai.com) supplies the core llms behind chat, analysis, and embeddings
• gemini (gemini.google.com) adds multimodal ai capabilities from google, from chat to vision tasks
• langfuse (langfuse.com) delivers llm observability, prompt versioning, and evals in one platform
• litellm (see oss-llmops-stack) acts as our unified llm proxy with model routing and failover
• langgraph (built on the langchain blog's langgraph platform) powers agent deployment at scale
• llmops (open-source reference at oss-llmops-stack.com) stitches everything into a cohesive, self‑hosted stack

open-source llmops stack

for end-to-end llm operations, we rely on the open source llmops stack—a ready-made combo of litellm (llm gateway) and langfuse (observability + prompt management), backed by vibrant oss communities

ready to dive in? follow the guides on oss-llmops-stack.com to deploy in minutes and take full control of your llm pipelines 🚀

Want to collaborate or have a question? Contact me.