On-Device AI Researcher — Fast, Tiny Models for Phones

Desert Ant Labs; Amsterdam; Remote; EUR 70,000 - 120,000; Full time

Job summary

Desert Ant Labs in Amsterdam offers a hands-on role to advance speech, vision, video, and text models for on-device deployment. You will help shrink models to fit real devices and ship them in our Detail and Subwave apps.

The role covers data generation, benchmarking across languages, and collaborating with teams to define model capabilities and evaluation standards. Remote work options exist alongside Amsterdam-based collaboration.

Qualifications

• Write solid Python and know PyTorch well.
• Ship models and explain end-to-end what was done.
• Write clean, tested code and review changes carefully.
• Experience with on-device runtimes and edge ML.

Responsibilities

• Generate synthetic and augmented data, test on real recordings, videos, and text from the devices people use.
• Measure latency, memory, heat, and battery use on real hardware before reporting a number.
• Build models that combine audio, video, and text, such as labeling who speaks when.
• Decide with Detail and Subwave what a model should do, then measure the model in the app.
• Write the model card and the release post for each model.

Skills

Python

PyTorch

Model benchmarking

Data generation

Tools

Core ML

LiteRT

ONNX Runtime

ExecuTorch

MLX

llama.cpp

coremltools

CUDA

Job description

Desert Ant Labs in Amsterdam offers a hands-on role to advance speech, vision, video, and text models for on-device deployment. You will help shrink models to fit real devices and ship them in our Detail and Subwave apps.

The role covers data generation, benchmarking across languages, and collaborating with teams to define model capabilities and evaluation standards. Remote work options exist alongside Amsterdam-based collaboration.