Posted 1h ago
Loading this job…
Applied Prompt Engineering Intern
Posted 2h ago0 views
Some of the best prompt engineers we have met started by fixing something small and shipping it.
The idea behind this one
This is a short, unpaid, fully remote internship built for students who want evidence they can do the work. Over eight weeks you will pull scoped prompt tickets off a shared board: a summariser that keeps dropping deadline dates, a classifier that confuses contract roles with freelance ones, a resume parser that mangles Indian address formats. Small problems, real users, visible fixes.
The loop you will repeat
Every ticket works the same way. Reproduce the failure with three examples, propose a prompt change, test it against a held out set, then write up what happened whether or not it worked. Negative results are genuinely welcome here and an honest one beats a vague success story.
Being upfront
There is no stipend attached to this position. It is built to fit around a final semester at roughly fifteen hours a week with no fixed daily hours. What you get is a mentor, merged work, and a written record of the fixes you shipped.
Responsibilities
- Pick up scoped prompt tickets from a shared board and carry each through to a merged fix
- Reproduce every reported LLM failure with at least three concrete examples before changing anything
- Test each prompt change against a held out example set
- Write a short result note per ticket, including the changes that did not work
- Improve the summarisation prompt so it stops omitting dates and application deadlines
- Refine classification prompts that confuse contract, freelance and full time roles
Requirements
- Final year student or recent graduate available for around fifteen hours a week
- Basic Python and comfort using an API key from a script
- Able to write a clear bug report with reproduction steps
- Honest about what you tried and what failed
- Working internet and availability for one weekly call
Skills
Benefits
- Certificate + Letter of Recommendation
- Weekly mentorship call with an applied AI engineer
- Flexible fifteen hour weeks designed to fit around college
- Your merged fixes credited by name in the internal changelog