Jobs of the Week (UK & Europe): Top 10 Applied AI & Agentic Systems Engineering Roles – Week of November 05, 2026
Curated weekly intelligence featuring the top 10 Applied AI, Agentic Systems, Sovereign RAG, and LLM Application Engineering roles across the UK and Europe with verified compensation (£/€/CHF), core AI stacks, and production use cases.
Weekly career intelligence for production-minded AI builders across the United Kingdom and Europe.
Compensation figures below are calibrated market benchmarks from the research cycle, not necessarily employer-published bands. Confirm live scope, location eligibility and pay directly with each employer before applying.
Executive briefing: Europe’s applied AI hiring market is becoming an infrastructure market
The strongest UK and European openings are no longer framed as generic “AI” jobs. They sit at the point where model capability meets production accountability: stateful agents that can call tools safely, retrieval systems that can cite and filter evidence, inference stacks that keep sensitive data within approved jurisdictions, and evaluation layers that make behaviour measurable rather than anecdotal.
London remains the region’s most aggressive cash-and-equity market for frontier deployment, fintech automation and developer tooling, with senior applied engineering packages commonly calibrated around £120k–£210k base before bonus or equity. Paris is increasingly defined by sovereign model deployment around Mistral-class open-weight systems; Berlin, Munich and Heidelberg combine industrial, public-sector and privacy-sensitive workloads with packages typically around €105k–€155k. Zurich remains an outlier for total compensation in specialist infrastructure and regulated-enterprise roles, where CHF 160k–240k is a realistic senior benchmark. Remote UK/EU roles expand access, but often retain country-specific payroll and compensation bands.
The architectural signal is clear. Employers want engineers who can choose between managed APIs and private inference, deploy Mistral, Mixtral or Llama-family weights through vLLM or TensorRT-LLM, connect those systems to Qdrant, Weaviate or pgvector, and then operate them with tracing, offline evaluation, access controls and human escalation. In regulated environments, the EU AI Act is moving from policy discussion to engineering work: risk classification, logging, data governance, transparency, robustness testing and evidence trails must be designed into the system—not appended after launch.
Fintech teams in London and Paris are concentrating on controlled, tool-verified automation for AML, disputes and customer operations. Munich and Zurich are extending the same principles into process intelligence, industrial workflows and enterprise systems of record. The best roles therefore reward a hybrid profile: a strong software engineer who can reason about model quality, systems reliability, security boundaries and user workflow design.
Top 10 comparison table
| Rank | Job title & company | Hub / arrangement | Calibrated compensation | Core applied AI & agentic stack | Sovereign enterprise mandate | Apply |
|---|---|---|---|---|---|---|
| 1 | Applied AI Engineer, Sovereign Institutions Mistral AI | Paris, hybrid | €115k–€160k + equity | Python, Mistral, vLLM, LangGraph, Qdrant, Kubernetes | Air-gapped and zero-retention deployments for government and financial institutions | Careers |
| 2 | Applied AI Architect, Enterprise & Startups Anthropic | London, hybrid | £130k–£200k + equity | Claude, Python/TypeScript, LangGraph, DSPy, Bedrock/Vertex | Governed enterprise agent workflows and evaluation systems | Careers |
| 3 | Applied AI Engineer, Codex & Technical Success OpenAI | London, hybrid | £140k–£210k + equity | OpenAI APIs, Python, TypeScript, FastAPI, Next.js, evals | Production code agents and verified enterprise tool use | Careers |
| 4 | Member of Engineering, Agent Sandboxes Poolside | Paris / remote EMEA | €140k–€210k + equity | Python, Rust, Firecracker, Docker, Kubernetes, vLLM | Isolated execution runtimes for self-verifying coding agents | Careers |
| 5 | Staff ML Engineer, Roleplay Sessions & Agents Platform Synthesia | London / UK-EU remote | £120k–£170k + equity | Python, PyTorch, WebRTC, LangGraph, Redis, Ray | Real-time multimodal enterprise learning and simulation agents | Careers |
| 6 | Lead Deployment Architect, Enterprise AI Celonis | Munich, hybrid | €110k–€155k + bonus/equity | Python, process intelligence graph, Databricks, SAP APIs, LLM orchestration | Grounded agents with safe ERP write-backs and auditable process context | Careers |
| 7 | Senior AI Solutions Engineer, Sovereign Systems Aleph Alpha | Heidelberg / Berlin, hybrid | €105k–€145k + options | Python, vLLM, Qdrant, on-prem Kubernetes, Slurm | Air-gapped public-sector, health and industrial GenAI | Careers |
| 8 | Staff GenAI Engineer, Autonomous Financial Agents Revolut | London, hybrid | £125k–£185k + equity/bonus | Python, LangGraph, LlamaIndex, vLLM, pgvector, Kafka | Deterministic financial-compliance and transaction workflows | Careers |
| 9 | Applied AI Systems Engineer, Autonomous Software Agents Cognition AI | London / remote UK-EU | £140k–£220k + equity | Python, TypeScript, Linux, tool harnesses, ASTs, test sandboxes | Long-horizon code-agent execution and deterministic validation | Careers |
| 10 | Staff Applied AI Engineer, Customer Operations & Multi-Agent RAG Klarna | Stockholm / Berlin / EU remote | €105k–€150k + equity | Python, DSPy, Weaviate/Qdrant, FastAPI, Terraform, observability | Multilingual customer-operation agents with strict factual controls | Careers |
In-depth profiles
1. Mistral AI — Applied AI Engineer, Sovereign Institutions
Initiative snapshot: Mistral sits at the centre of Europe’s open-weight and sovereign AI push. This profile supports institutions that need modern LLM capability without surrendering data residency, deployment control or auditability.
Compensation & benefits: Calibrated at €115k–€160k base plus meaningful BSPCE-style equity; seniority and customer-facing deployment depth should materially influence the offer.
Stack: Python, Mistral model APIs and weights, vLLM or TensorRT-LLM, LangGraph, Qdrant/Weaviate, containers and Kubernetes. Evaluation, access-control and observability experience is particularly valuable.
Production challenge: Delivering multilingual RAG and agent systems for public-sector and financial customers with private data boundaries, citation requirements and predictable structured outputs.
Ideal candidate: A senior builder who has taken LLM applications from prototype to secure production, can discuss GPU inference trade-offs, and understands EU procurement and high-risk-system evidence expectations.
Application: Mistral jobs portal · Mistral careers
2. Anthropic — Applied AI Architect, Enterprise & Startups
Initiative snapshot: Anthropic’s London deployment work connects frontier Claude capabilities with enterprise operating models, including high-assurance tool use and evaluation-led implementation.
Compensation & benefits: £130k–£200k base plus equity and comprehensive benefits, with premium positioning for senior technical advisory and architecture ownership.
Stack: Python, TypeScript, Claude tool use, LangGraph, LlamaIndex, DSPy, AWS Bedrock or Google Vertex AI, and Ragas/TruLens-style evaluation harnesses.
Production challenge: Designing agent workflows that are useful without being unconstrained: permissions, approval gates, retrieval quality, prompt/version management and regression testing all matter.
Ideal candidate: A technically credible architect who can build alongside customer teams, explain system trade-offs to executive stakeholders and establish measurable quality gates.
Application: Anthropic careers
3. OpenAI — Applied AI Engineer, Codex & Technical Success
Initiative snapshot: OpenAI’s London applied engineering work focuses on turning frontier developer and reasoning capabilities into reliable enterprise products.
Compensation & benefits: £140k–£210k base plus equity and benefits; highly competitive for candidates who combine full-stack delivery with customer-facing technical leadership.
Stack: Python, TypeScript, Node.js, FastAPI, Next.js, structured outputs, function calling, containerisation and automated eval suites.
Production challenge: Building code copilots and document-analysis systems where agents plan, use tools, inspect outputs and pass programmatic checks before an action reaches a user or system of record.
Ideal candidate: A senior software engineer with practical LLM product delivery, robust API design and a rigorous test-first approach to agent behaviour.
Application: OpenAI jobs portal · OpenAI careers
4. Poolside — Member of Engineering, Agent Sandboxes
Initiative snapshot: Poolside’s engineering opportunity is an infrastructure role disguised as an AI role: autonomous coding systems need safe, fast and reproducible places to execute untrusted code.
Compensation & benefits: €140k–€210k base plus early-stage equity, reflecting scarce expertise across systems engineering and agent evaluation.
Stack: Python, Rust or Go, Firecracker microVMs, Docker, Linux cgroups/eBPF, Kubernetes, Git internals, Tree-sitter and distributed inference services.
Production challenge: Enabling thousands of isolated compile-test-debug loops without cross-tenant leakage, runaway resource use or non-reproducible validation.
Ideal candidate: A systems engineer with deep Linux and isolation knowledge who is comfortable linking runtime reliability to code-agent quality.
Application: Poolside jobs portal · Poolside careers
5. Synthesia — Staff ML Engineer, Roleplay Sessions & Agents Platform
Initiative snapshot: Synthesia is evolving enterprise video into interactive, responsive training and communication experiences where speech, avatar rendering and LLM orchestration must remain coherent in real time.
Compensation & benefits: £120k–£170k base plus equity and flexible benefits; remote eligibility depends on the specific role and payroll country.
Stack: Python, PyTorch, C++ where latency demands it, WebRTC, streaming APIs, LangGraph, Redis, Ray and inference-serving infrastructure.
Production challenge: Managing context, interruption, latency and safety across live conversations while integrating enterprise learning or support systems.
Ideal candidate: A staff-level ML/platform engineer with real-time systems experience and strong product instincts around multimodal interactions.
Application: Synthesia jobs portal · Synthesia careers
6. Celonis — Lead Deployment Architect, Enterprise AI
Initiative snapshot: Celonis brings a differentiator that many agent deployments lack: a detailed view of how work actually moves through business processes and systems of record.
Compensation & benefits: €110k–€155k base plus performance bonus and equity, with the strongest packages reserved for architects able to lead complex client transformations.
Stack: Python, process intelligence graphs, LangChain/LlamaIndex patterns, Databricks, PySpark, REST interfaces and SAP/Salesforce connectors.
Production challenge: Grounding recommendations in process data, then safely writing actions back into ERP systems with controls, traceability and escalation paths.
Ideal candidate: A deployment architect who understands both enterprise integration and LLM system design—and can make automation legible to operational owners.
Application: Celonis careers
7. Aleph Alpha — Senior AI Solutions Engineer, Sovereign Systems
Initiative snapshot: Aleph Alpha’s positioning is built around European control, transparency and deployment in environments that cannot freely export data to external model providers.
Compensation & benefits: €105k–€145k base plus stock options, with upside for engineers experienced in on-premise and air-gapped customer environments.
Stack: Python, foundation-model serving, vLLM, Qdrant/OpenSearch, Kubernetes, bare-metal GPU estates, Slurm and confidential-computing patterns.
Production challenge: Building explainable, citation-led RAG systems for public administration, defence, healthcare and manufacturing while satisfying strict security controls.
Ideal candidate: An enterprise solutions engineer who is as comfortable with identity, networking and deployment constraints as with prompt design and retrieval evaluation.
Application: Aleph Alpha careers
8. Revolut — Staff GenAI Engineer, Autonomous Financial Agents
Initiative snapshot: Revolut represents the regulated-fintech version of agentic engineering: high event volume, stringent controls and limited tolerance for opaque automation.
Compensation & benefits: £125k–£185k base plus equity and performance bonus, with an emphasis on senior operators capable of owning production reliability.
Stack: Python, Java/Kotlin-adjacent services, LangGraph, LlamaIndex, vLLM, pgvector, Kafka, Redis, Kubernetes and cloud data platforms.
Production challenge: Supporting AML investigations, disputes and service triage with deterministic tool verification, human review and full decision traceability.
Ideal candidate: A staff engineer who understands event-driven systems, model risk and financial-services control environments.
Application: Revolut careers
9. Cognition AI — Applied AI Systems Engineer, Autonomous Software Agents
Initiative snapshot: Cognition’s autonomous software-agent work raises the bar from autocomplete to long-horizon task execution across repositories, tools and tests.
Compensation & benefits: £140k–£220k base plus significant venture equity, reflecting the premium for rare agent-runtime and developer-tooling expertise.
Stack: Python, TypeScript, Linux internals, AST tooling, Git, vector retrieval, containerised execution and test harnesses.
Production challenge: An agent must inspect a codebase, select actions, produce patches and prove those changes work inside an isolated environment—without confusing plausible output for validated output.
Ideal candidate: A strong systems or developer-tools engineer with exceptional debugging instincts and a bias for rigorous automated verification.
Application: Cognition jobs portal · Cognition careers
10. Klarna — Staff Applied AI Engineer, Customer Operations & Multi-Agent RAG
Initiative snapshot: Klarna is a practical reference point for high-volume, multilingual AI customer operations, where business value only appears if the system stays accurate, fast and appropriately constrained.
Compensation & benefits: €105k–€150k base plus RSU-style equity; location, seniority and scope will affect the final package.
Stack: Python, OpenAI/Anthropic APIs where appropriate, DSPy, Weaviate or Qdrant, FastAPI, Terraform, cloud infrastructure and Langfuse-style observability.
Production challenge: Handling financial customer journeys across languages while grounding responses in current policy, routing edge cases and monitoring factual or operational failures.
Ideal candidate: A full-stack applied AI engineer who can connect retrieval quality, API reliability, observability and human-support operations into one production system.
Application: Klarna careers · Klarna jobs portal
Strategic takeaways: an upskilling roadmap for UK and European builders
- Build stateful agents, not prompt demos. Learn explicit state graphs, idempotent tool calls, retries, approval gates and durable checkpoints. LangGraph-style orchestration is valuable because it makes agent behaviour inspectable.
- Treat evaluation as a product surface. Create versioned test sets, trace failures, measure retrieval relevance and faithfulness, and add deterministic assertions around tools and structured outputs. Ragas, TruLens, DeepEval and custom harnesses are useful only when connected to real failure modes.
- Become fluent in sovereign deployment choices. Practice serving open weights with vLLM, model routing, GPU capacity planning, vector filtering, tenant isolation and secure networking. Be able to explain when an external API is appropriate and when residency or confidentiality requires a private stack.
- Translate the EU AI Act into engineering controls. Develop habits around risk documentation, data lineage, logging, transparency notices, human oversight, security testing and post-deployment monitoring. Regulated employers need implementers, not merely policy awareness.
- Keep one foot in software engineering. The most durable roles favour engineers who can ship APIs, queues, schemas, observability, integration tests and operational runbooks alongside model workflows.
Application strategy: Present portfolio evidence as production case studies: the business workflow, architecture diagram, model and retrieval choices, guardrails, evaluation results, latency/cost constraints, and the operational response when a tool call or retrieval path fails. That evidence maps directly to the hiring signal across this week’s list.
Research references: employer career portals linked above. Listings and availability can change quickly; check the official portal before submitting an application.