AI Research Engineer (London)

Edra is solving one of the hardest problems in enterprise AI: AI models are generic but company processes are specific. We build AI agents that learn how processes actually run, and then run their operations.

We're a Series A startup, backed by Sequoia and other leading VC firms, led by Co-Founders who created and led Forward Deployed AI Engineering at Palantir, and we're growing our team in New York and London. We're a deeply technical team of engineers, AI researchers, and strategists with a high bar for talent and a shared belief that exceptional people are the foundation of everything great we'll build.

We're building a learning system that teaches AI agents how enterprises actually work. Our system starts from what a customer already has: knowledge base articles, support conversations, tickets, system logs, and agent traces. From that evidence it reverse-engineers how the business actually runs and writes the process down as instructions a person can inspect and an agent can execute, with confidence scoring to know when to automate and when to seek human input.

We're looking for AI Research Engineers who have built complex, production LLM-based systems. Whether you've scaled LLM workflows handling millions of requests, built multi-agent systems in production, or designed evaluation frameworks for enterprise deployments, we want people who bring intensity and self-direction to their craft. You'll work on our agentic learning environment: the loops that turn human corrections and agent outcomes into better knowledge and instructions over time. The job combines applied research with ownership of production systems. You find a failure, form a hypothesis, run the experiment, and ship the fix into real customer workflows.

You'll spend most of your time advancing our core context-learning library and the harness that powers our agents across every customer. The work spans continuous learning, context engineering, agent orchestration, and evaluation. You'll contribute to our knowledge benchmarks and to the self-improving capabilities of the harness.

• →

  Build and extend our core context-learning library, turning customer-informed problems into reusable platform capabilities.

• →

  Design and implement LLM-powered systems and agentic workflows from concept to production.

• →

  Build agentic features for knowledge management: agents that can edit, update, and maintain large knowledge bases autonomously.

• →

  Build continuous learning loops that use human corrections and production outcomes to find knowledge gaps, detect changed procedures, and improve prompts and instructions.

• →

  Build reliability and confidence systems: evaluation frameworks, confidence scoring, and logic for when to automate and when to escalate to a human.

• →

  Architect async, scalable systems that handle complex AI orchestration.

• →

  Contribute to the direction of our AI strategy and product capabilities.

• You've built a complex, production LLM system with real depth: multiple layers of engineering decisions you can walk through, including what changed after real users arrived.

• You have practical depth in one or more of context engineering, retrieval, agent orchestration, memory, evaluation, or learning from feedback.

• You have experimental judgment. You can turn an ambiguous problem into a testable hypothesis, set a baseline, and tell a real improvement from noise.

• You're excited by open-ended problems where the solution might not exist yet.

• You can navigate between research papers and production code with equal comfort.

• You're a strong communicator who can explain your work to teammates and customers.

• You have 3+ years of professional experience.

Don’t meet all of the above traits? We’d still love to hear from you. We’re looking for exceptional people with unique skills and interests, and we know that great people have different backgrounds and skillsets.

• Watch our Series A announcement video

• Read our blog post from our launch

• Learn more about what we're building

• Our founders point of view on the market

We work together in person five days a week from our offices New York and London.

Edra is an equal opportunity employer and we encourage applications from individuals of all backgrounds. Edra does not discriminate on the basis of race, color, religion, sex, national origin, age, disability, genetic information, marital status, sexual orientation, gender identity, veteran status, or any other legally protected characteristic.