What You'll Do
Design and refine multimodal large language models to expand product capabilities and enhance real-world deployments across cloud and edge environments. Develop novel neural network architectures that deliver high accuracy while operating efficiently on low-power hardware. Optimize training pipelines and improve data processing workflows to accelerate model development cycles.
Run distributed training jobs using PyTorch and TensorFlow on Kubernetes-based infrastructure. Leverage GCP-hosted GPUs to train models and automate data labeling at scale. Use Snowflake and Dataflow to construct robust, scalable datasets. Rapidly prototype new features and interactive demos using Streamlit to showcase technical advancements.
Requirements
- Demonstrated software engineering skills with strong proficiency in Python
- Hands-on experience with deep learning frameworks such as PyTorch, TensorFlow, Keras, or JAX
- Proven background in computer vision and working with multimodal LLMs
- Track record of training neural networks that have transitioned into production systems
Preferred Qualifications
- Industry experience deploying models efficiently on edge devices or cloud platforms
- Background in Deep Reinforcement Learning applications
Benefits
- Competitive salary and generous equity package
- Relocation support available
- Choose your preferred laptop and development equipment
- 25 days paid vacation plus public holidays
- Opportunity to attend leading research conferences including NeurIPS, ICML, and CVPR
Work Environment
This is a hybrid role based in London or Amsterdam. You’ll have flexible scheduling with the expectation to work from the office at least two fixed days per week. The technical stack includes Python, PyTorch, TensorFlow, JAX, Kubernetes, Snowflake, Dataflow, Streamlit, and GCP GPU resources.
