Senior AI Product Engineer

What you will do

• Design and build agentic workflows and AI-powered features that power Elliptic's copilot, using LLM frameworks such as LangChain or LangGraph.
• Lead the design of agents, tool integrations, and retrieval pipelines that turn complex blockchain data into useful answers for investigators.
• Build and own evaluation frameworks that measure and improve the quality, reliability, and latency of LLM outputs.
• Design and build the backend services, APIs, and event-driven systems that support these features, using TypeScript and Node.js.
• Drive technical design reviews and architecture decisions for AI workstreams.
• Mentor and coach more junior engineers through pair programming, code review, and design feedback.
• Raise the engineering bar across the team by promoting good practices in testing, observability, and AI system reliability.
• Influence cross-team decisions on how AI capabilities integrate with the rest of the Elliptic platform.

What you will achieve in the first 6 months

• Owned and shipped at least one significant AI capability end-to-end, from ambiguous problem statement through to production.
• Driven the technical direction of an AI workstream, including architecture, evaluation, and rollout.
• Established or improved at least one team practice for building AI systems (evals, observability patterns, prompt management, rollout safety).
• Mentored junior engineers on AI engineering practices and contributed to their growth.
• Built strong working relationships across product, web engineering, data engineering, and DevOps.
• Started to influence adjacent teams on how Elliptic builds with AI.

You will be a great fit here if you

• Care about building AI systems that work reliably in production, not just demos.
• Have strong instincts for trade-offs specific to LLM systems: cost vs latency vs quality, deterministic vs probabilistic behaviour, when to use an agent vs a workflow.
• Can take ambiguous AI problems and break them into shippable pieces with clear evaluation criteria.
• Communicate clearly with non-engineers (product, stakeholders, leadership) about AI capabilities and limitations.
• Have coached or mentored other engineers and enjoy raising the bar through reviews, pairing, and feedback.
• Are genuinely curious about the domain and find blockchain investigations and financial crime prevention interesting.

Our ideal candidate has

• 5-8+ years of software engineering experience, with at least 1-2 years of meaningful production work building LLM-powered features.
• Direct experience with LLM frameworks such as LangChain, LangGraph, or similar.
• Hands-on experience building agentic systems: tool use, multi-step reasoning, planning, memory, and human-in-the-loop patterns.
• Practical understanding of prompt engineering, structured outputs, context management, and managing the trade-offs of working with LLMs.
• Experience building and running evaluations for LLM outputs (eval sets, LLM-as-judge, regression testing).
• Strong backend skills in TypeScript / Node.js, with solid API design.
• Cloud experience (AWS: Lambda, ECS, S3, or similar).
• Database proficiency across SQL (Postgres) and some NoSQL exposure.
• Demonstrated ability to mentor and coach other engineers.
• Uses AI coding assistants (Copilot, Cursor, Claude, etc.) critically and effectively as part of their day-to-day workflow.

Bonus points for

• Experience with retrieval-augmented generation (RAG) and vector databases (pgvector, Pinecone, or similar).
• Experience designing multi-agent systems and agent orchestration patterns.
• Familiarity with LLM observability tools (LangSmith, Langfuse, Arize, or similar).
• Experience working with multiple model providers (Anthropic, OpenAI, open-weight models) or with fine-tuning.
• Hands-on experience with Terraform, Kubernetes, or infrastructure-as-code tooling.
• Experience with observability platforms like Datadog.
• Distributed or event-driven architectures (SNS, SQS).
• Interest in cryptocurrency, blockchain, or compliance.