Posted 11h ago
LLM Applications Engineering Intern
Posted 2d agoBe an early applicant
Six months, one product surface
MailerMen's outreach assistant drafts the first message a founder sends to a candidate, and right now it writes like a form letter. You will own the retrieval and prompting work that makes it sound like the person actually sending it.
What you will actually build
You build a retrieval layer over company profiles, role descriptions and previously accepted messages, using sentence embeddings and a vector index, then run structured prompt experiments against it. Expect to write an offline evaluation harness before you write a single prompt, because the team will ask you for pass rates rather than impressions. Across the six months you also implement guardrails for hallucinated company facts, add a caching layer for repeated context, and instrument the whole assistant so drafting quality can be tracked week over week.
Who tends to do well here
This one is open worldwide and runs fully asynchronously, so you should write clearly and be able to unblock yourself. Final year students and fresh graduates are welcome, and no professional experience is expected. Familiarity with Python and at least one LLM API from something you built for yourself matters far more to us than a course certificate does.
Responsibilities
- Build a retrieval layer over company profiles, role descriptions and previously accepted outreach messages
- Design and run structured prompt experiments with results recorded, not remembered
- Write an offline evaluation harness before shipping any prompt change
- Implement guardrails that catch hallucinated company facts before a message is sent
- Add a caching layer for repeated context and measure its effect on latency
- Instrument drafting quality so it can be tracked week over week
- Document prompt versions so a teammate can reproduce your results
Requirements
- Final year student or recent graduate anywhere in the world, no work experience required
- Python and comfort calling an LLM API from your own project
- Basic understanding of embeddings and vector similarity search
- Clear written English for asynchronous collaboration across time zones
- Available for a six month internship
Skills
Benefits
- Competitive monthly stipend paid in INR, open to candidates worldwide
- Fully asynchronous remote setup with no fixed office hours
- Direct mentorship from the applied AI lead
- Certificate of completion and a LinkedIn recommendation
- Priority consideration for a remote full time role