Principal Engineer, CoCounsel
On-siteZug, Zug, Switzerland
Job Summary
Define patterns for building MCP servers, agents, and scalable infrastructure that turn frontier models into production-grade workflows for legal professionals. Build and evolve systems operating over millions of documents, privileged data, and thousands of concurrent AI interactions while optimizing cloud costs. Lead initiatives across AI, product, identity, and infrastructure; mentor staff and raise the bar on integration practices and security. Design identity and access strategies including SSO, RBAC/ABAC, and least-privilege service identities to ensure safe autonomous agent actions. Establish observability, incident response, and SLOs for high-throughput, low-latency workloads handling complex law and retrieval pipelines.
Required Qualifications
- Bachelor's Degree in Computer Science, Computer Engineering, a related field, or equivalent experience
- Demonstrated experience building with AI – LLMs, agents, and retrieval
- Proven track record owning large, complex projects end-to-end: architecture, execution, rollout, and long-term operation
- Deep Python expertise and experience with production systems using frameworks like FastAPI (or similar)
- Experience with relational databases (PostgreSQL or equivalent)
- Experience with a major cloud provider (AWS preferred)
- Strong background in distributed systems: data modeling, API contracts, observability, resilience patterns, and performance tuning under load
- Hands-on experience with identity and access management – designing or integrating authentication and authorization at scale (SSO, SAML, OIDC, OAuth 2.0, RBAC/ABAC, token/session management, or comparable IAM systems)
Desired Qualifications
- Hands-on experience designing and operating agentic systems in production – tool calling, MCP servers, multi-step workflows, and orchestration around third-party and in-house LLMs (e.g., Anthropic, OpenAI), including prompt/response management, cost controls, and safety considerations
- Experience with AI-adjacent infrastructure: vector databases, embeddings, semantic search, or custom retrieval pipelines
- Opinions and experience around automated testing, reliability, and release practices for systems with non-deterministic model behavior
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.