Thesis: Hybrid agentic workflow analysis

As AI-assisted development becomes more common, cloud model usage limits and costs create a need for alternative workflows. Increasingly capable laptops make it possible to run smaller open-source language models locally. This thesis will explore how local and cloud-based models can cooperate in agentic development to reduce reliance on cloud resources and lower costs.

The assignment

• Investigate a hybrid agentic development workflow in which a larger language model supervises and delegates tasks to smaller models running locally.
• Propose a solution using an agent harness such as OpenCode, with delegation of tasks including code exploration and implementation.
• Assess whether smaller models can successfully carry out tasks assigned by larger models.
• Identify the hardware a developer would need and evaluate the complexity of setting up the workflow.
• Examine whether the approach requires particular agent harnesses or tailored system prompts.

Who we are looking for

• Studying towards a bachelor's or master's degree in Computer Science or Engineering, with graduation planned for June 2027.
• Fluent in both spoken and written Swedish and English.

Practical details

Based at Knightec Group's Sundsvall office. Start in January or March 2027. Apply through the company's careers page as soon as possible, and by 30 November 2026 at the latest. Questions can be sent to johanna.edstrom@knightec.se.