5 Month FTC - up to £200 per day - London/Hybrid About Us: We're an AI Ed-Tech startup using our AI software to help teachers with formative assessment in schools saving teachers time and helping with retention of staff to save schools recruitment costs.
Project Overview: Drive the technical development, deployment, training, and testing of an AI-based formative assessment system that leverages computer vision, NLP, and machine learning to automate GCSE mathematics marking and deliver pedagogical feedback and insights for teachers.
This project includes working with advanced AI models such as YOLOv8, Llama, and SAM, implementing FASTAPI, and emphasizing test-driven development-spanning unit tests, end-to-end testing, and custom benchmarks to ensure high standards of pedagogical accuracy and quality.
Key Responsibilities: Lead AI/ML development, training, inference deployment and system architecture Carry out model distillation and synthetic data generation to achieve benchmarks target from limited dataset Make a RLHF reward model to improve feedback quality of the insights and use it to train Llama model that can run on desktop hardware.
Build a retrieval chatbot over the database and deploy it through an email interface to provide planning insights Required Skills: Experience with PyTorch, TensorFlow, or any modern AI frameworks Worked with and built computer vision and NLP systems Proficiency in Python, FASTAPI, and microservices architecture, Experience with local model optimisation and deployment Open to all kinds of applicants ideally 2-3 years experience but open to a junior who has a desire to learn with or without CS qualifications, we look for people who have built interesting projects on their own and or with a team.