SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Ledgy is building a modern, tech-driven equity management and financial reporting platform for private and public companies across Europe. The company raised a $22M Series B led by New Enterprise Associates in 2022 and operates from offices in Zurich and London.
You'll join the AI Systems team to build the AI foundations and product experiences of Ledgy, including the Ledgy Agent (an in-app AI assistant for equity administrators), the Ledgy MCP server (enabling customers to bring equity data into their existing AI tools), and the platform capabilities that let other teams embed AI safely across the product.
Key responsibilities include:
**Ledgy Agent**: Design and build the system behind the in-app agent that enables equity administrators to perform tasks through AI that they currently do manually. The agent takes real actions on live equity data (e.g., stakeholder onboarding, analysis) and must be accurate and observable. You'll design the tools, skills, context, orchestration, and decision-making logic.
**Ledgy MCP**: Shape the MCP server that lets customers bring equity into the AI agents and tools they already use. You'll ensure capabilities built once work in both the in-app agent and MCP, and decide what to expose while maintaining safety.
**Measurement and Improvement**: Design and build systems to measure agent quality and learn from customer usage through evaluations, tracing, and feedback from real conversations. You'll define what "good" means, how to measure it, and how insights drive future development.
**Scaling**: Design foundations that let AI features be built once and reused everywhere. You'll handle backend infrastructure, cost, latency, reliability, and observability as usage grows.
You'll own real parts of the system, have input on investment priorities, and spend time with customers to understand how administrators use the agent and inform what to build next.
The tech stack is TypeScript-based: React and Tailwind on the front end; tRPC, Node.js, MongoDB, and Temporal on the back end. The agent is built on the Vercel AI SDK with an MCP server alongside it. Infrastructure runs on GCP with Kubernetes and Terraform.
Location: Hybrid for London-based candidates (2 days in office per week). Remote options available across the UK, Spain, Germany, Poland, and Romania.
**Requirements:**
- 5+ years of experience as a Software Engineer with strong TypeScript/Node fundamentals across back end and front end
- Shipped agentic or LLM-based software to production (agent, assistant, chat product) and operated it in production, debugging failures and managing cost/latency
- Designed agents: tools and skills, context and prompt engineering, state, streaming, agent loop; deep knowledge of at least one agent framework or runtime
- Use evidence to improve systems: evals, regression sets, tracing, and real usage feedback to validate changes
- Stay current with foundation model labs and community learning on agent building; evaluate and adopt new techniques early
- Participate in product decisions: read data, talk to customers, advocate for priorities
- Hands-on experience building with LLMs strongly preferred
