Research Intern: Low-Latency Speech & LLM Inference

AGIGO; Zürich; Vor Ort; CHF 20.000 - 29.000; Vollzeit

Benefits dieser Stelle

Paid internship

Flexible hours

Access to GPU clusters (Hopper, Blackw

Zusammenfassung

AGIGO invites Master’s or PhD students (preferred) or recent graduates in CS/ML to join a 6-month internship focused on low-latency, high-quality voice AI infrastructure.

You will build quantization pipelines, experiment with 4-bit weight representations, KF-cache quantisation, and sparsity while ensuring all variants meet a universal speech-quality gate across Hopper/Blackwell/ADA hardware.

Qualifikationen

• Master’s or PhD student (preferred) or recent graduate in Computer Science, Machine Learning, or a related field
• Strong Python programming skills and Git
• Solid understanding of ML fundamentals and MLOps
• Hands-on experience with PyTorch
• Fluent in English, highly motivated, willingness to learn

Aufgaben

• Develop quantization-aware training pipelines for internal LLMs and TTS models
• Implement quantization, pruning, and hardware-adaptive build variants
• Build and maintain a quality gate with metrics like UTMOS, WER, and speaker similarity
• Explore elastic serving and speculative decoding for multiple hardware targets
• Collaborate with researchers and engineers to translate research into deployable optimizations

Kenntnisse

Python programming

Git

ML fundamentals

PyTorch

English fluency

Ausbildung

Master’s or PhD student (preferred) in CS/ML

Recent graduate in CS/ML

Tools

N/A

Jobbeschreibung

AGIGO invites Master’s or PhD students (preferred) or recent graduates in CS/ML to join a 6-month internship focused on low-latency, high-quality voice AI infrastructure.

You will build quantization pipelines, experiment with 4-bit weight representations, KF-cache quantisation, and sparsity while ensuring all variants meet a universal speech-quality gate across Hopper/Blackwell/ADA hardware.