You will work with Python, TensorFlow or PyTorch, and large language models, building custom deep-learning models and training/fine-tuning LLMs on a proprietary dataset. Kangal AI builds frontier AI for governments, national security, and private corporations. As a founding hire, you will own the model stack and scale the research into a team you lead.

Work with Python, containerized pipelines, and cheminformatics representations (SMILES, InChIKey, molecular fingerprints) to build ML evaluation benchmarks for molecular structure prediction from mass spectrometry data. You will design adversarial checks, leakage-safe datasets, and repeatable pipelines that measure model quality for an applied AI drug discovery company.

Fitch Group★4.5 (4)· London

Service Reliability Engineers embed with Fitch Ratings development squads, using AWS, Azure, Kubernetes, Docker, GitHub Actions, Datadog, and AWS Bedrock/SageMaker. They drive reliability, observability, and AI-enabled operations for mission-critical services, focusing on new AI development and cloud infrastructure automation.

Work with C++ and Python on multithreaded, real-time trading infrastructure. You will develop ultra-low-latency exchange protocol encoders and fine-tune fully automated trading strategies for a data-driven prop fund. The role is in a small, fast-paced team that collaborates with researchers and operations.

They will use Python, LLM tooling (prompting, RAG, agents, MCP, fine-tuning, evaluation), and AI platforms/APIs. Working on PIMCO's Tactical AI Automations team, they deploy AI assistants and agentic workflows across Portfolio Management and Analytics, and support Lane Assist/

You'll lead a 6-7 person team owning core ML infrastructure—distributed model training, LLM hosting/fine-tuning, and scalable deployment—using Python heavily with working C++/Java, at a top systematic quant fund. This hands-on engineering manager role drives ML into high-stakes production systems.

Architect low-latency inference pipelines for multimodal LLMs, TTS, and Vision models using Python, C/C++, PyTorch/JAX, TensorRT/ONNX, and custom CUDA/Triton kernels. Optimize with Speculative Decoding and KV-Cache quantization. Iconic Interactive is a seed-stage AI-native video game studio building intelligent virtual actors that speak, move, and react in real-time.

You will work with Python, FastAPI, TypeScript, React, and GCP to design and build services, data models, and APIs. Electric Twin builds a behavioral simulation platform that creates "digital twins" of populations to help organizations predict human behavior.

You'll work with LLMs, chain-of-thought prompting, RAG architectures, multi-agent tool use, Python, and FastAPI. Electric Twin builds synthetic populations for behavioral prediction, simulating human responses for global brands using LLMs.

You will develop and deploy production-grade LLM-powered agentic systems with persistent memory, tool orchestration, and RAG pipelines, integrating with Google Ads, Meta, and CRMs. You will build autonomous agents for marketing operations at a Series A startup building an AI-native teammate for marketing teams. As the second ML hire, you will shape the production stack and product roadmap.

Wells Fargo★3.6 (48k)· London

You'll work with Python, SQL, cloud and DevOps practices, AI/ML and Large Language Models, plus Axiom ControllerView on the Enterprise Functions Technology team. The role enhances and supports regulatory reporting platforms used by Wells Fargo's Finance Controllers organization.

Canva★4.3 (32)· London

You'll work with Python, PyTorch, vision-language models (Gemma, Qwen-VL, SmolVLM, MiniCPM-V), optimization techniques (quantization, pruning, distillation, hardware-specific compilation, task-specific fine-tuning), and on-device runtimes (Core ML, LiteRT/TFLite, ONNX Runtime, ExecuTorch). The Video Storytelling team builds models and systems behind Canva’s video AI, partnering with the Edge AI group to explore on-device deployment. This internship optimizes a video-capable VLM for on-device intelligent captioning on consumer phones.

You will work across the full model lifecycle, training and fine-tuning PyTorch-based diffusion/DiT models and using diffusers, transformers, torch.distributed, torchrun, Accelerate, and PEFT/LoRA, with Hugging Face datasets and ffmpeg for data. The team builds and scales generative video models for VEED's AI video creation platform, most recently releasing Fabric, a DiT-based speech-to-video model.

Design, build and deploy AI solutions using Python, Large Language Models (LLMs), orchestration frameworks, CI/CD, containerisation, and MLOps/LLMOps. You will join a newly established AI function, shaping its future by developing scalable, production-ready AI applications from concept to deployment.

You'll work with Python, PyTorch, SGLang, LLMs, embeddings/reranking models, PEFT/LoRA, RAG, Docker, microservices, and cloud infrastructure. The team builds AI-powered applications transforming specialized financial markets, developing and deploying LLM solutions with significant autonomy. You'll join a small, experienced engineering team collaborating directly with founders and domain experts.

Capgemini★3.7 (11k)· London

Works with Neo4j, Amazon Neptune, Azure AI Search, Elasticsearch, Pinecone, Python, Azure, AWS, Kubernetes, Docker, and CI/CD. Designs and implements an enterprise knowledge platform using knowledge graphs, RAG, and GraphRAG to enable AI agents across Insurance, Payments, and Healthcare domains. Leads platform strategy and cross-functional engineering teams.

Capgemini★3.7 (11k)· London

Designs and implements enterprise knowledge platforms for AI agents across Insurance, Payments, and Healthcare domains, using graph databases (Neo4j, Amazon Neptune, TigerGraph, ArangoDB), retrieval technologies (Azure AI Search, Elasticsearch, Pinecone, Weaviate, Qdrant), and Python on cloud-native stacks (Azure, AWS, Kubernetes, Docker).

You will work with SQL, Python, AWS or Azure, Snowflake

Work with Python, TensorFlow, PyTorch, scikit-learn, FastAPI/Flask, Docker, Kubernetes, open-source LLMs, vector databases, and AWS/Azure/GCP. Design, build, and deploy ML models, develop Generative AI and RAG pipelines, and fine-tune LLMs for a health-focused AI team.

Luxoft★4.5 (669)· London

You'll build an investment-insights solution using Agentic AI, working with Python, FastAPI, AWS, Git/CI/CD, vector databases, RAG patterns, and LLMs like OpenAI/Anthropic. The project generates investment insights from sales and research materials to speed client-meeting preparation. Nice-to-have skills include React/TypeScript and AgentCore, but team size is unstated.

You'll work with Python, modern ML tooling, LLMs, and simulations, owning the AI/ML stack including model selection, fine-tuning, and inference infrastructure. Prior Foundry builds specialized AI agents and inference systems to solve complex global policy problems for the public sector in social science and public policy.

You’ll use Python, Docker, Kubernetes, CI/CD tools, cloud platforms (AWS, Azure, GCP), and IaC tools like Terraform or Ansible, with AI frameworks such as PyTorch or TensorFlow. As an Applied AI Engineer at Mistral, you’ll onboard customers and deploy production AI solutions across infrastructure, back-end, and front-end. No team size is stated.

This role focuses on building integrations between xAI's Grok AI models and international government systems using Python, APIs, SDKs, and cloud platforms. The team engineers and deploys AI capabilities for policy analysis, education, public health, and citizen services across Americas, Asia, and the Middle East, requiring hands-on work on model fine-tuning, safety evaluation, compliance, and real-world stakeholder collaboration. You'll need 4+ years of software engineering experience, prior AI model fine-tuning in government contexts, and expertise deploying AI systems in sovereign environments with international regulatory constraints.