Artificial Intelligence Intern

Role overview

We are looking for a Software Intern for a three-month internship to design and run benchmark measurements of artificial intelligence models used in engineering work. The role focuses on benchmark design, repeatable measurement, accurate reporting of results, and documentation thorough enough that another engineer can reproduce the measurements later. It is a self-contained software project rather than firmware or satellite hardware work, and its output is working code and a written report rather than a publication. Reporting to a senior engineer.

Location: Satlabs Facility in IIT Madras Research Park, Chennai.

Required qualifications

• Currently enrolled in, or recently graduated from, a degree in any engineering or computing discipline.
• University coursework covering the fundamentals of artificial intelligence or machine learning.
• Working proficiency in Python, sufficient to build and maintain a small tool independently.
• Working comfort with the Linux command line.
• Ability to write clearly for a mixed technical and business audience, including short reports intended to support a decision.
• A measurement-led approach to technical questions, with attention to the validity of reported numbers.

Preferred experience

• Hands-on work with machine learning models or their programming interfaces beyond coursework, in personal projects, research, or industry.
• Benchmarking, measurement or experimental work of any kind, including outside software.
• Familiarity with Git, databases, and containers.
• Interest in satellites or space systems.

Key competencies

• Accuracy in reporting results, including clear statements of which results are reliable and which are not.
• Systematic approach to problem-solving, with clear documentation of method and findings.
• Ability to work independently, with the judgement to seek direction when it is needed.
• Willingness to raise blockers early, which matters particularly over a twelve-week engagement.
• Attention to the distinction between what was measured and what has been inferred from it.

Screening and evaluation process

The selection process is short and is intended to assess both technical capability and the measurement discipline the project depends on.

• Application review: Assess evidence of work built or measured and described accurately, whether a repository, an academic project, or a written report.
• Introductory call: Short discussion of the role, the project, availability, and the candidate's objectives.
• Technical assessment: Time-boxed exercise in Python, involving reading unfamiliar code, making a small change to it, running it, and reporting the result accurately.
• Technical conversation: Discussion of a project the candidate has built or measured, covering method, results, and the limits they would place on their conclusions. Coursework and personal projects are equally acceptable.
• Final discussion: Role expectations, supervision arrangements, compensation, what the internship leads to, and alignment with the engineering culture.