Brief: Develop an explainable assistant for Lustre storage configuration within the HPC Cluster Configurator, covering MDS/MDT, OSS/OST, clients, capacity/bandwidth and workload suitability.

Goals and tasks:

• Define structured inputs (clients, capacity, bandwidth target, metadata intensity, file size profile, workload type).
• Build a simplified configuration model, add validation rules, and recommend suitable options.
• Generate explanations for recommended/risky/invalid configurations; create test datasets; document methodology and limits.

Profile/skills:

• Python; knowledge of AI/LLM prompting or explainable recommendation; OS/storage basics; interest in HPC; strong writing.
• Plus: Lustre/parallel FS knowledge, throughput/capacity/striping concepts, REST APIs, basic frontend skills.

Duration & perks:

• Target: final-year entrants; candidates interested in a future PFE are encouraged.
• Period: 2–3 months.
• Paid: Monthly stipend.