We’re partnering with Gistr to find an AI Engineer.

This is not a conventional “build an LLM feature” role.

Gistr is building an AI-native notebook for learning, research and thinking — bringing together things people consume online and turning them into knowledge they can actually work with. Today, that spans videos, PDFs, podcasts, AI-assisted research and notes. Gistr is already used by 15,000+ learners, students and educators.

The team is now looking for an engineer who wants to work much closer to the frontier of agentic systems.

You’ll be building and shipping production AI agents end-to-end — not just prototypes.

The work includes

→ Designing AI agent systems with LangGraph and/or DeepAgents

→ Context engineering — deciding what information, tools and memory an agent should have at each step

→ Building agent harnesses: tools, control flow, guardrails, retries and evaluation loops

→ Instrumenting and improving agents with LangSmith or comparable observability tooling

→ Building production backend services in Python and FastAPI

→ Working with streaming APIs and WebSockets

→ Containerising and deploying services with Docker

→ Taking problems from “this should exist” to something running in production

→ Moving across backend, infrastructure, data and product when the problem requires it

What matters here isn't just whether you know the stack.

They’re looking for someone who has actually built and shipped AI agents, can reason about why an architecture should work, researches before committing to an implementation, and is comfortable owning ambiguous problems end-to-end.

The must-haves are deliberately specific:





- Hands-on experience building and shipping AI agents
- LangGraph and/or DeepAgents experience
- Practical context engineering experience
- Experience with agent harnesses
- LangSmith or comparable LLM observability experience
- Solid Python + FastAPI experience
- Production experience with streaming APIs and WebSockets
- Docker

AWS or comparable cloud infrastructure experience is a plus.

This is a role for someone who would rather show us something they built than tell us they are “passionate about AI.”

And the hiring process reflects that.

1. Initial conversation
2. Technical discussion around your actual agent-building experience
3. Hands-on engineering exercise
4. Deep dive into your implementation, tradeoffs and reasoning
5. Final conversation with the team

The exercise is intentionally designed to look simpler than it is. The goal isn't simply to see whether you can make code work. It's to understand how you approach the problem, what you choose not to build, the tradeoffs you make, and how you reason about the system. If you’re interested, don’t send us a keyword-stuffed resume and call it a day.

Send

- Resume
- GitHub or equivalent
- 2–3 projects you’re genuinely proud of
- Anything you’re currently building, even if unfinished

Side projects, open source and work you can share all count.

If some of your best work is under NDA, just tell us. Don’t leave the answer blank.

Charlie is partnering with Gistr on this search, and we’re specifically looking for people whose actual work matches the problem — not people who happen to match a list of keywords.

Remote · Full-time

📌 AI Engineer - Gistr - Full time - Remote (India)
🏢 Charlie
📍 India