Computer Vision Researcher - London (Perm)

microTECH Global Limited; Greater London; Hybrid; GBP 90,000 - 120,000; Full time

Job summary

microTECH Global Limited in London is pursuing a permanent, full-time role focusing on frontier research in multimodal AI and large language models. You will design ViT-based architectures, advance data pipelines, and develop scalable training systems for real-world deployments.

The role emphasizes strong Python/PyTorch skills, deep learning expertise, and collaboration with cross-functional teams to translate research into production-ready solutions.

Qualifications

• Bachelor’s degree or higher in Computer Science, Mathematics, Statistics, or related technical field.
• Strong Python experience with PyTorch and DL frameworks.
• Excellent algorithm design, mathematical reasoning, and problem solving.
• Strong communication and collaboration across cross-functional teams.

Responsibilities

• Develop ViT and multimodal large model architectures with improved reasoning and efficiency.
• Advance multimodal alignment, representation learning, and long-context modelling.
• Explore scalable training methods for large multimodal models.
• Optimize model architectures for generalization and performance.
• Integrate multimodal capabilities into production and user-facing applications.

Skills

Python

PyTorch

Deep learning

Algorithms

Communication

Cross-functional teamwork

Mathematics

Education

BSc or higher in CS/Math/Stats

Job description

Backed by an international research team and abundant computing resources, the centre focuses on core research directions including multimodal understanding and generation, vision-language large models, and embodied intelligence.

This is a permanent, full-time position located in the tech hub of King's Cross, London.

Key Responsibilities:

Frontier Technical Breakthroughs

• Develop ViT and multimodal large model architectures with improved reasoning and efficiency
• Advance multimodal alignment, representation learning, and long-context modelling
• Explore scalable training methods for large multimodal models
• Optimize model architectures for generalization and performance

Data Ecosystem Construction

• Process large-scale multimodal data across images, videos, audio, and text
• Build pipelines for data cleaning, filtering, annotation, and quality control
• Construct and maintain datasets with versioning and reproducibility
• Optimize data mixtures and sampling strategies for model training
• Improve data quality through feedback-driven curation loops

MLLM Systems & Infrastructure

• Build distributed training systems for large-scale multimodal models
• Optimize GPU utilization, cluster efficiency, and resource scheduling
• Develop open-source training frameworks for scalable model development
• Engineer training, inference, and serving infrastructure
• Improve scalability, stability, and performance of model systems

Business Value Delivery

• Integrate multimodal capabilities into assistant and content generation scenarios
• Translate research into production and user-facing applications
• Collaborate with product and engineering teams to deploy and iterate models

Person Specification:

Essential Requirements:

• Academic Background: Bachelor’s degree or above in Computer Science, Mathematics, Statistics, or related technical disciplines.
• Technical Skills: Proficient in Python programming with strong hands-on experience in PyTorch and deep learning frameworks.
• Core Competencies: Strong algorithm development and implementation skills, solid mathematical and logical reasoning ability, and excellent cross-functional communication and collaboration skills.
• Traits: Self-driven and highly motivated toward advancing artificial intelligence (AI), with strong resilience and the ability to tackle challenging technical problems.

Desired:

• Strong track record of publications in top-tier AI or computer vision conferences (CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR)
• Hands-on experience in large-scale model pre-training or fine-tuning
• High-impact open-source projects or internship experience in leading technology companies within CV, NLP, or multimodal domains