GTA opportunity – Django development and deployment for Feedback Loop (apply by 15 Oct)
We are seeking one or two PhD students from the Department of Computing and/or the Department of Electrical and Electronic Engineering to help complete and evaluate Feedback Loop, a web application that supports GTAs and course leads in improving student feedback. This is a paid, part-time opportunity to develop an existing application and contribute to an education project using large language models (LLMs).
Appointment details
Duration: An agreed period ending no later than 31 July 2027.
Hours: Flexible part-time hours to be agreed within the project budget, with a maximum of 20 hours per week per student, including any other GTA work.
Pay: Tier 1 £20.84 per hour plus £2.52 holiday pay (£23.36 total), or Tier 2 £25.35 plus £3.06 holiday pay (£28.41 total), subject to confirmation of the appropriate appointment category and tier.
The project
The project, Transforming GTA Feedback with LLMs: Improving Clarity, Consistency, and Actionability at Scale, aims to make feedback clearer, more actionable and aligned with assessment rubrics. Feedback Loop already provides course and assignment management, marking workflows, grade moderation and CSV export. The LLM integration and production deployment remain to be completed. Academic staff and markers retain responsibility for assessment decisions and final feedback.
Main responsibilities
- Take over the existing Python and Django codebase, review the documentation and agree priorities with the project team.
- Implement LLM API integration and prompts for feedback improvement and moderation, using assignment specifications and rubrics while preserving human review.
- Prepare and deploy the application on a Linux server, including production configuration, database arrangements, static and uploaded files, logging and backups.
- Work with relevant College teams on authentication and single sign-on, and on VLE integration or suitable data exchange where agreed.
- Test permissions, marking workflows and AI outputs; improve the interface; support a pilot with GTAs and teaching staff; document deployment, maintenance and handover.
Essential skills and experience
- Current PhD student with availability for an agreed part-time commitment through the appointment period.
- Practical Python and Django experience, including models, forms, views, templates and authentication; ability to understand and extend an existing codebase.
- Experience deploying web applications on Linux, using the command line, managing dependencies and configuring production services.
- Experience integrating APIs, handling JSON responses, errors and credentials securely.
- Working knowledge of relational databases, SQL and Django ORM migrations, plus HTML, CSS and JavaScript for interface changes.
- Git version control, debugging and automated testing, with attention to permissions and secure handling of student assessment data.
- Ability to work independently, communicate progress clearly and produce useful technical documentation.
Desirable skills and experience
- LLM API integration, prompt design and evaluation of AI outputs for accuracy, usefulness, bias and unsupported claims.
- Production Django hosting with tools such as Nginx and Gunicorn, PostgreSQL, HTTPS and service management.
- Single sign-on integration using OpenID Connect, OAuth or SAML, and experience with Canvas or other virtual learning environment APIs.
- Extraction of text from assignment specifications and rubrics in PDF or Word formats.
- GTA marking experience, assessment and feedback practice, accessibility, user testing or qualitative evaluation.
How to apply
Please send a short CV and a brief statement outlining your relevant experience, your availability and preferred weekly hours. Include links to relevant code or projects where available, and explain your contribution to any team projects.
Contact: Dr Zohaib Akhtar, Dr Iro Ntonia
Applications to: z.akhtar@imperial.ac.uk, i.ntonia@imperial.ac.uk
Closing date: 15-10-2026