This position is offered through the ANU Computing Internship courses (COMP4820 / COMP8830).
Semester 2, 2026 applications open on Monday 18th May 2026 and close on Sunday 31st May 2026.
Company
Clairva.ai is a Singapore-incorporated startup building licensed video and audio data infrastructure for multimodal AI training. We source content from global media libraries and commission specialised video data collection, then produce annotated datasets that meet Tier-1 ingestion standards for foundation-model labs and dataset platforms. We prioritise rights-cleared provenance and high-quality annotation pipelines.
Project
Clairva.ai is a Singapore startup working with licensed video data for multimodal AI training. Multi-object tracking on egocentric video is an open research area. State-of-the-art methods shifted in late 2024 with SAM-2-based trackers. There is no clean public benchmark comparing classic and SAM-2 era trackers on egocentric data. Schema-driven validation of video annotations is also an active topic.
The student investigates open tracking methods on public datasets. They learn benchmarking methodology and modern computer-vision practice. They explore schema-driven evaluation of video annotations.
The student will benchmark four to five open tracking models. Examples are ByteTrack, BoT-SORT, StrongSORT, SAMURAI, and MASA. Datasets are Ego4D and EPIC-Kitchens-100. Metrics include HOTA, MOTA, IDF1, and runtime.
The student will also design a JSON Schema for annotated egocentric clips. They will implement a Python validator and a set of test fixtures. They will build a small CLI tool that ties tracker output to the schema.
Deliverables fit into four work packages.
- The first is the tracker benchmark report and reference script.
- The second is the schema, validator, and 20 test fixtures.
- The third is the CLI tool on at least 10 public clips.
- The fourth is a written research report and a 30-minute presentation.
Required technical skills
Required skills are Python 3.10+ with NumPy and PyTorch basics. Computer-vision basics are needed, including object detection, tracking, and video frame I/O.
A Git and GitHub workflow with pull requests and code review is expected.
The intern should set up open-source ML repos and run inference scripts.
Preferred skills are prior coursework in multi-modal ML, computer vision, or video analysis. Familiarity with the HuggingFace ecosystem is a plus. Basic exposure to JSON Schema also helps.
Required/preferred professional and other skills
Required skills are clear written technical communication. The intern should produce README-quality docs and a short report. They should be willing to read recent ML papers with supervisor guidance. Independent problem-solving on tractable engineering tasks is expected. Comfort working remotely with a weekly sync cadence is needed. Preferred skills are prior open-source contribution.
Delivery Mode
Hybrid
Student location
Project’s Special Requirements/ Conditions
None
Type of internship
Unpaid
How to apply
Applications are invited from eligible students to apply for the Computing Internship courses COMP4820 or COMP8830. Eligibility details of COMP4820 / COMP8830 and further information about the Computing Internship can be found on the Computing Internship page.
Eligible students can apply through the Computing Internship application form which will be available via the Computing Internship page between Monday 18th May 2026 and close on Sunday 31st May 2026.
You can nominate multiple preferred Internship projects and host organisations through the one online application form.
Eligibility and room available in degree to undertake COMP4820/COMP8830 will be assessed at the time of application. If you do not meet the eligibility criteria or do not have room in your degree to fit COMP4820/COMP8830, your application will not be progressed.
Your application will require you to upload the following documents:
- an updated copy of your resume, and
- an expression of interest (limit 350 words) for each organisation you wish to apply to (organisations with multiple projects may only submit one expression of interest, so state clearly which project/s you wish to be considered for).