Karya has a live employer posting for a Research Engineering Intern in AI Evaluations, based in Bengaluru, India. The role focuses on building and running evaluations that measure how AI models work across languages and contexts, rather than simply demonstrating a chatbot.
The posting does not state a salary or stipend, application deadline or internship duration. Those details should be checked directly with the employer; this article does not fill them in from assumptions or third-party estimates.
Who the employer is looking for
- Applicants currently pursuing, or recently having completed, a bachelor’s or master’s degree in Computer Science, Machine Learning, NLP or a related field.
- Strong Python skills and comfort working with data, using pandas or similar tools.
- Working knowledge of LLMs and modern machine learning, including familiarity with calling models through APIs.
- A solid foundation in probability and statistics, with methodical debugging skills.
Karya lists experience with evaluation, annotation or human-feedback data as a differentiator. Exposure to speech or multilingual NLP, especially Indian languages, and experience with GPU inference, cloud tools or Android/app debugging are also differentiators. They should not be confused with a blanket requirement to have every listed skill.
What the work involves
The intern will help prepare datasets, run model inference through APIs and self-hosted setups, and configure annotation tasks on Karya’s platform. The role also includes monitoring live evaluations, reproducing app issues, coordinating fixes and documenting recurring problems.
On the analysis side, the posting describes automated data checks, identifying anomalies, computing agreement and scores, and comparing automated judges with human ratings. It also includes building tested, documented tooling with logged configurations so that evaluation work can be reproduced.
That makes this an engineering-and-research operations role. A relevant project example would explain how data was checked, how a model was evaluated and how a failure was debugged, rather than only listing tools used. This is application-preparation guidance, not an additional employer requirement or a promise of selection.
What to confirm before accepting an offer
Ask the employer about pay, duration, start date, working arrangements and the selection process. Bengaluru is the location in the posting; no remote-work entitlement should be assumed. A live listing also does not guarantee that a vacancy will remain open indefinitely.
Use only the Karya employer application page on Greenhouse to review the current role and apply. JobAt360 is reporting the opportunity, not collecting applications or offering placement. Do not pay anyone for a guaranteed selection.
Original summary of the employer posting, checked on November 19, 2026. The employer’s current page controls eligibility, availability and application instructions. No salary, deadline or duration is supplied because none is stated there.
For more English and Hindi explainers, visit the free Wolf News app.
