Audio | Multimodal ML Engineer

We’re hiring an Audio / Multimodal ML Engineer for a fast-growing AI infrastructure startup building the safety and control layer for large-scale AI systems.

• AI-native product company operating in the AI safety & infrastructure space

• Backed by international investors

• Processing large-scale AI traffic across enterprise environments

• Training and fine-tuning proprietary models for performance & reliability

• Small, highly technical team shipping fast

The company builds the control and evaluation layer for AI systems - helping organizations define, test, and enforce how AI behaves in real-world environments.

• Train and fine-tune large-scale audio & multimodal models

• Design and run experiments (architecture, data mixtures, training strategies)

• Build and optimize audio data pipelines

• Improve inference speed, latency and production readiness

• Deploy models end-to-end in low-latency environments

• Define meaningful evaluation metrics beyond benchmark scores

• Collaborate closely with research & engineering

This is a hands-on role where research meets production.

• PyTorch-based training pipelines

• Large-scale distributed training

• Speech & audio modeling architectures

• Multimodal model integration

• Model optimization (quantization, distillation, streaming inference)

• Production deployment & serving systems

(Deep technical stack shared during interviews)

• 3+ years training deep learning models in audio / speech domains

• Strong experience with distributed training frameworks

• Solid understanding of audio signal processing fundamentals

• Experience shipping models to production (latency matters, not just metrics)

• Experience building and maintaining data pipelines

• Strong engineering hygiene (clean code, testing, versioning)

• Experience with multimodal architectures

• Experience with alignment / fine-tuning techniques

• Experience working in AI infrastructure or model optimization