Python, PyTorch, JAX, LLMs, vLLM, vector databases and modern agent frameworks will be used. The stealth AI startup is building a pre-launch, next-generation AI assistant for everyday tasks, conversations and workflows — a 0-to-1 opportunity on a small engineering team.

This role develops generative AI applications using Java, Python, React or Angular, LangChain, LlamaIndex, Hugging Face, Docker, Kubernetes, and AWS, with tools like GitHub Copilot. You will build scalable, AI-powered solutions for Citi’s global finance operations.

On Ag★2.9 (7)· London

Work with Python, PyTorch, JAX, TensorFlow, GCP/AWS, Kafka, and MLOps to build real-time personalization and recommendation engines. Lead data science initiatives that design, train, and deploy customer-level algorithms for On’s global D2C/e-commerce platform.

You'll work with PyTorch, distributed training, Docker/K8s/IaC, and Python backend. The company builds foundation models for extreme physics to accelerate progress in semiconductors, aerospace, defence, and fusion. Joining a small, research-heavy team, you'll be the first dedicated engineering owner for the ML stack.

DNEG★3.1 (36)· London

Design, build, and maintain high-performance RESTful APIs, microservices, and event-driven architectures using Python, PostgreSQL, and Redis. You will integrate generative AI models (ATMAN, VAANI) into production backends for a company operating at the intersection of enterprise Media Asset Management and generative media.

You will work with Python, C#, JavaScript, or Go, and use Gen AI tools like LangGraph, CrewAI, and Hugging Face, alongside ML frameworks such as PyTorch. Sitting in a cross-functional team with product, data, and platform engineers, you will design, build, deploy, and monitor AI systems that scale across the business.

Revolut★3.4 (183)· London

You will work on Python, PyTorch, TensorFlow, Keras, OpenCV, Jupyter, Docker, and GCP, with SQL and CI/CD. The team builds and deploys computer vision models to automate compliance processes like ID verification and facial recognition for account onboarding.

Citi★3.9 (21k)· London

Build and scale AI-powered financial systems using Java, Spring, React/Angular, Python, FastAPI, Docker, and GitHub Copilot within Citi’s global tech environment. You will design generative AI agents, LLM workflows, and RAG pipelines while working on enterprise microservices, mobile, or front-end development. This role is for an individual contributor on an agile team.

You’ll build multimodal 3D models fusing dental scans with clinical text using Open3D, trimesh, PyTorch3D, PyTorch Geometric, YOLO, Mask R-CNN, CNNs, transformers, and point/graph networks. The role owns end-to-end research on converting oral scans into staged treatment plans for 01Health’s multi-specialty healthcare platform.

Fitch Ratings★3.8 (136)· London

You'll build ML/AI solutions using Python, FastAPI, PyTorch/TensorFlow, LangChain, LangGraph, and React/TypeScript within cloud-based microservices architectures. The agile team delivers software products for Fitch's BMI business, which provides data-driven market insights and risk analysis to customers globally. You'll design full-stack enterprise applications, implement GenAI features with LLM integration, and establish development best practices across the product suite.

Work with Python, Kubernetes, vLLM, SGLang, and cloud infrastructure (AWS, Azure, GCP) to build POCs, MVPs, and production integrations. As an AI Field Engineer, you embed with customers to deploy and fine-tune LLMs (SFT, DPO, RFT) and architect production inference systems.

You will work with Python, PyTorch, JAX, LLMs, vLLM, vector databases and modern agent frameworks. The team builds a pre-launch, next-generation AI assistant for managing everyday tasks, conversations and workflows. This is a small, stealth startup.

Build B2B biology products that put foundation models into scientists’ hands. Work across the stack with Python, TypeScript/React, SQL, REST/gRPC APIs, AWS, Docker, Kubernetes, CI/CD, and AI coding tools (opencode, Claude).

Checkout.com★3.8 (15)· London

Build systems for training, deploying, and monitoring ML models on Databricks using Python, AWS, Azure, and frameworks such as scikit-learn, xgboost, TensorFlow, PyTorch, and Spark. You will contribute to Checkout.com’s new ML-driven dispute optimisation suite, building data pipelines, feature stores, and backend services for real-time payment dispute processing.

Work with Python, C++, and machine-learning frameworks (TensorFlow, PyTorch) on cloud platforms (AWS, Azure, GCP) using Docker and Kubernetes. You will bridge the gap between AI research and product development, building scalable software that integrates advanced mathematical-reasoning algorithms into real-world applications.

Work with Python, SQL, and ML stacks (PyTorch, TRL) to build evaluation and data pipelines for LLMs and AI coding agents. The team owns how AI agents generate, analyze, and improve Kotlin code across Android, KMP, and server platforms, creating benchmarks and post-training pipelines.

Citi★3.9 (21k)· London

You will work with Java, Angular, microservices, event-driven architectures (Kafka, RabbitMQ), Spring Boot, Docker, Kubernetes, and AI frameworks (Python, LangChain). At Citi, a global financial services firm, you will design and deploy AI-enabled full-stack applications, integrating agentic AI to solve complex business challenges.

Autodesk★4.1 (589)· London

Use Python, CI/CD (GitHub Actions, Jenkins), and AI-powered testing tools to build and maintain automated test suites for AI-powered applications. This role ensures the quality, reliability, and performance of Autodesk’s AI features through API testing and validation of machine learning model outputs.

Funding Circle★3.5 (53)· London

You will work with React Native, React.js, TypeScript, Firebase, Gradle, XCode, REST APIs, and CI/CD for app releases on iOS and Android. The role involves building and scaling a mobile app for Funding Circle’s fintech platform that uses data to provide fast funding to small businesses.

You'll build and maintain backend services and APIs for Zuma Labs' OTC trading platforms using Python, Django, PyTorch, and SGLang, collaborating with frontend engineers, product teams, and traders. The company develops force multipliers for systemic industries, with products targeting over-the-counter trading markets. You'll design scalable microservices, optimize performance, and solve complex real-world problems with broad autonomy and direct impact on the business.

Encord is building a universal data layer for AI that helps teams train and run models, handling indexing, curation, annotation, and evaluation across the AI lifecycle. You'll work with PostgreSQL, ClickHouse, Python, Kubernetes, and GCP to build highly performant backend systems managing billions of structured and unstructured data items, taking end-to-end ownership of projects from architecture through deployment and impact measurement.

Encord is building a universal data layer for AI that helps teams train and run models on the right data. You'll work full-stack with Python, TypeScript, React, Kubernetes, and GCP, building and extending foundational systems across backend services and frontend interfaces while owning projects end-to-end from architecture through deployment and impact measurement. The role emphasizes autonomy, problem-solving at scale, and collaboration within a small, high-trust team.

Encord builds a data platform for AI teams to index, curate, annotate, and evaluate training data across the AI lifecycle. As a Product Engineer, you'll own features end-to-end using Python, TypeScript, React, Kubernetes, and GCP, working on backend systems, ML workflows including model evaluation and active learning, and polished frontends for annotation teams and ML engineers. You'll ship rapidly while collaborating with customers and product to deliver high-impact solutions.

You'll build high-performance frontend experiences for Encord's Physical AI data platform, working with TypeScript, React, WebGL, and Python across the full stack—from 3D rendering and browser optimization to backend data pipelines and storage systems on GCP/Kubernetes. The platform helps AI teams index, curate, and annotate data for robotics and autonomous vehicle applications at scale.

You'll build large-scale backend infrastructure in Python and TypeScript/React on Encord's data platform that helps AI teams train and run models, deployed on Kubernetes and GCP. The role involves end-to-end ownership from architecture to deployment, with experience in prompt engineering, agent orchestration, and AI tool integration valued. You'll join a 60-person team at a crucial growth stage, working autonomously on complex scaling challenges with direct customer impact.