Master Thesis – Core Tech Media (Audio AI)

Location: Lund, Sweden; remote and on-site workplace options. The thesis is open only to students affiliated with a Swedish university or college, either directly or through an exchange program.

Employment: Full-time, fixed-term thesis worker. Scope: two Master students, 30 hp each.

Company

hirify.global develops network technology and security solutions covering video surveillance, access control, intercom, audio systems, and intelligent analytics.

What you will do

• Choose and research one of four thesis topics: knowledge distillation for audio classification; evaluation of generated and simulated training data; energy-efficient audio classification for battery-powered edge devices; or lightweight speaker anonymization for intercoms.
• Develop and evaluate audio classification or anonymization methods for camera, intercom, embedded, or IoT security products.
• Measure practical trade-offs including classification accuracy, real-world generalization, energy consumption, latency, memory, compute requirements, privacy, and speech intelligibility.
• Build reproducible pipelines, benchmarks, prototypes, or design guidelines based on the selected research topic.
• For speaker anonymization, address real-time full-duplex operation, narrowband or wideband audio, codec artifacts, outdoor noise, reverberation, and informed-attacker robustness.

Requirements

• Current affiliation with a Swedish university or college, directly or through an exchange program.
• Master-level studies in Computer Science, Applied Mathematics, Technical Physics, Electronics Engineering, or a similar field.
• Relevant coursework in C or Python and Machine Learning or AI.
• For the speaker anonymization topic, relevant coursework in Signal Processing.
• Ability to work on a 30 hp Master thesis; proposals involving two students require applying in pairs and identifying the co-applicant.
• CV and university or college grade summary must be submitted with the application.

Nice to have

• Experience with audio classification, sound event detection, synthetic or simulated audio data, generative audio models, embedded hardware, model compression, quantization, pruning, knowledge distillation, signal processing, speech technology, or speaker verification.

Culture & Benefits

• Work on security, intercom, audio, and intelligent analytics products used to improve safety and business performance.
• Collaborative environment with colleagues and partners across more than 50 countries.
• Applications are accepted in Swedish or English.
• Background checks may apply to certain roles, with notice provided before action is taken.

Hiring process

• Apply through the proposal advert in Swedish or English.
• Submit a CV and university or college grade summary.
• For two-student proposals, submit separate applications and clearly name the co-applicant.