AI Engineer Intern

Aspect; Greater London; On-site; GBP 15,000 - 20,000; Full time

Job summary

Aspect, one of London's largest property maintenance teams, seeks an AI Engineer Intern to work on real-world computer vision and multimodal AI projects within the Data & AI team. You will gain hands-on experience with modern vision models, embeddings, and production‑oriented software development.

You will research, prototype, and evaluate approaches using Python, PyTorch, OpenCV, and AI APIs, handling image and video data with guidance from mentors.

Qualifications

• Fundamentals of ML and deep learning are expected.
• Experience with computer vision through projects or internships is desirable.
• Familiarity with PyTorch, TensorFlow, OpenCV, scikit-learn or Hugging Face.

Responsibilities

• Research and prototype AI solutions for computer vision and multimodal tasks.
• Build and evaluate models using Python, PyTorch, OpenCV, and AI APIs.
• Work with image/video data, including preprocessing and labeling.
• Experiment with image classification, object detection, segmentation, OCR.
• Develop evaluation frameworks for accuracy, latency and cost.
• Collaborate with cross-functional teams to translate problems into AI solutions.

Skills

ML fundamentals

CV basics

Tools

PyTorch

TensorFlow

OpenCV

scikit-learn

Hugging Face

Job description

At Aspect, we are one of London’s largest property maintenance teams, covering more trades than almost anyone else and operating 24/7.

After more than 25 years of serving thousands of residential and commercial customers, we are entering an exciting phase of digital transformation. We are investing heavily in AI, computer vision, automation, and data to rethink how property maintenance is delivered, from diagnosing problems and supporting engineers in the field to improving customer experience and operational efficiency.

We are looking for a talented and curious AI Engineer Intern to join our Data & AI team and work on real-world Computer Vision and multimodal AI projects.

This is not an internship where you will spend your time preparing presentations or working on artificial exercises. You will work with real data, real operational problems, and modern AI technologies, with the opportunity to build solutions that could ultimately be deployed across the business.

The Role

As an AI Engineer Intern, you will work closely with our Data & AI team to research, prototype, evaluate, and build AI solutions focused primarily on Computer Vision and multimodal AI.

You could be working on problems such as identifying property defects from images, analysing photos and videos from field engineers, extracting information from visual documents, matching images to jobs or assets, or combining images with text and operational data to help automate diagnosis and decision-making.

You will gain hands‑on experience with modern vision models, Vision‑Language Models, deep learning, model evaluation, AI APIs, and production‑oriented software development.

We are looking for someone who enjoys experimenting, building, and learning quickly. You do not need to know everything already, but you should be comfortable with Python, understand the fundamentals of machine learning, and be excited about turning new AI research into practical applications.

Example Projects

Depending on business priorities, you may work on projects such as:

• Building Computer Vision models to identify property defects, damage, equipment, materials, and maintenance issues from photographs.
• Using Vision‑Language Models to understand and describe images captured by customers and field engineers.
• Developing AI‑assisted property diagnosis tools that combine images, customer descriptions, and historical job data.
• Experimenting with object detection, image classification, segmentation, OCR, visual search, and image similarity.
• Building systems to automatically assess the quality and completeness of engineer photographs.
• Identifying objects, tools, components, boilers, plumbing fixtures, electrical equipment, or building materials from images.
• Creating multimodal AI applications combining vision, text, audio, and structured operational data.
• Evaluating state‑of‑the‑art models from OpenAI, Google, Anthropic, Meta, Hugging Face, and the wider open‑source AI ecosystem.
• Exploring synthetic data generation and data augmentation to improve model performance.
• Building prototypes that move beyond notebooks and become usable internal applications or APIs.

You Will

• Research, prototype, and evaluate Computer Vision and multimodal AI approaches for real operational problems.
• Build models and applications using Python, PyTorch, OpenCV, Hugging Face, and modern AI APIs.
• Work with image and video datasets, including data cleaning, labelling, preprocessing, augmentation, and analysis.
• Experiment with image classification, object detection, segmentation, OCR, embeddings, and Vision‑Language Models.
• Develop evaluation frameworks to measure model accuracy, reliability, latency, and cost.
• Compare different models and approaches and clearly communicate the trade-offs between them.
• Help turn successful experiments into APIs, tools, and production‑ready applications.
• Work with cloud AI platforms and GPU‑based development environments.
• Collaborate with Data, Product, Operations, and field teams to understand real‑world problems and translate them into AI solutions.
• Document experiments, findings, model performance, and technical decisions.
• Keep up with developments in Computer Vision, multimodal AI, foundation models, and applied machine learning.
• Contribute ideas and challenge existing approaches, we want interns who are curious and willing to experiment.

What We Are Looking For

We care more about your ability to learn, build, and solve problems than having years of commercial experience.

You should have:

• A good understanding of machine learning and deep learning fundamentals.
• Some experience with Computer Vision through university projects, personal projects, research, competitions, or previous internships.
• Familiarity with frameworks such as PyTorch, TensorFlow, OpenCV, scikit-learn, or Hugging Face.
• Understanding of concepts such as convolutional neural networks, transformers, embeddings, classification, object detection, or segm