Research Engineering Intern, AI Evaluations

Karya’s mission is to leverage the AI revolution to create dignified earning and learning opportunities for low-income communities across the world. In India, our smartphone-based digital work platform has enabled over 120,000 people to build language datasets and evaluate AI models while earning fair wages. With pilots in Kenya and Ethiopia and partnerships with leading AI labs, we are advancing language technologies for underrepresented communities. By combining dignified digital work with frontier AI innovation, Karya hopes to build a new AI economy that is equitable, inclusive, and community driven.

About the Role:

Karya's AI Evaluations team measures how well AI models serve people in the languages and contexts global benchmarks miss. We work with partners including Anthropic and Microsoft Research, alongside government and civil society, running evaluations on our own platform with native speakers and domain experts across India and beyond. Our benchmarks, leaderboards and reports aim to help model developers and decision-makers understand where models work well and where they fall short.

As a Research Engineering Intern, you will work alongside research engineers to build and run these evaluations, from setting up annotation tasks and running model inference to analysing results. You will own real parts of the pipeline end to end, including keeping them running when things break.

What You'll Do

• Evaluation setup: prepare datasets, run model inference through APIs and self-hosted setups, and configure, batch and launch annotation tasks on the Karya platform.
• Pipeline operations: monitor live evaluations, triage and reproduce annotation app issues, coordinate fixes with platform and operations teams, and document recurring problems.
• Quality control: run automated checks on annotation data and flag anomalies before they affect results.
• Analysis and experiments: compute agreement, scores and breakdowns; support focused studies such as validating automated judges against human ratings; and contribute to technical reports and public releases.
• Infrastructure: build and maintain tested, documented tooling for data processing, QC and analysis, with logged configs for reproducibility.

What You Bring

Must-Have:

• Currently pursuing or recently completed a Bachelor's or Master's in Computer Science, Machine Learning, NLP or a related field.
• Strong Python and comfort working with data (pandas or similar).
• Working knowledge of LLMs and modern ML, and familiarity with calling models via APIs.
• Solid foundation in probability and statistics.
• Methodical debugging skills, and willingness to dig into unfamiliar systems when something breaks.

Strong differentiators

• Experience with ML projects involving evaluation, annotation or human feedback data.
• Exposure to speech or multilingual NLP, especially Indian languages.
• Experience with GPU inference, cloud tooling or Android/app debugging.

Mindset

• Takes ownership of work end to end, knowing that reliable operations are what make evaluations trustworthy.
• Thrives in interdisciplinary teams spanning engineering, research, social science, operations and domain experts.
• Comfortable with ambiguity, messy data and shifting priorities.
• Cares about building AI that works for the people it is meant to serve.

Karya celebrates diversity and is an equal opportunity employer. All applicants will be considered without regard to race, religion, gender identity, sexual orientation, disability, or any other protected status.