This isn’t some resume-padding internship.
At PixelPai, we’re building compliance-grade Web3 infrastructure that actually works- modular, licensed, and ready for the onchain world.
We’re looking for a curious, technically sharp student who wants to help us push the limits of what AI agents can do.
We don’t expect you to be perfect. We expect you to show up hungry and willing to drown a bit before you learn to swim.
You’ll get thrown into the deep end of the pool 
But you’ll also build real, production-level AI agents using bleeding-edge tools like:
– LangChain, AutoGen, CrewAI, and Model Context Protocol (MCP)
– You’ll learn prompt engineering (few-shot, CoT, all of it)
– You’ll wire up persistent memory with vector DBs (Pinecone, Chroma, etc.)
– You’ll build observability + tracing with tools like LangSmith, Phoenix, or Weights & Biases
– You’ll design eval frameworks that actually test performance and reliability
We’re talking real-world work – debugging agent logic, failing forward, shipping weird ideas, and making them smarter every week.
You might be right if you:
– Write solid Python – or can challenge Claude, ChatGPT, or any AI out there to write it for you (but let’s be honest… you still need a rough idea of what you’re doing 
)
– Aren’t afraid of words like RAG, embeddings, or multi-agent coordination (or you’re ready to Google them hard)
– Think tracing, testing, and optimization are just as fun as launching cool demos
– Want to build things that don’t just live in notebooks
What you get:
– A front-row seat in a future unicorn
– Zero bureaucracy, 100% creative chaos
– Hands-on learning from a team that’s still figuring things out- but moves fast
– Flexible hours, real responsibility, and a chance to punch way above your weight class
To Apply:
Send us at career@pixelpai.com your:
LinkedIn or CV, GitHub, side project, or proof you build cool sh*t
3 lines on why AI agents get you fired up
We don’t care about age, degrees, or titles.
We care if you can think, build, and push.
