What you will do
Architect the "Unified Remediation Engine". You will design the high-throughput decision logic that powers our remediation efforts. You will build the backend services that intelligently route requests between deterministic instant QuickFixes, LLM-based reasoning fixes, and complex Agentic flows. Build Robust Agent Architecture. You will contribute to the Sonar Remediation Agent code, transitioning research concepts to a resilient, high-scale production service. You will implement self-healing workflows where the Agent iteratively fixes the code it produces until the build is green. Solve the "Sandboxing" Problem. You will architect the secure execution environments required to run untrusted, AI-generated code. You will design the isolation layers that allow Agents to iterate on code improvements without compromising our infrastructure. Engineer for Reliability & Safety. You will implement critical safety mechanisms, like Circuit Breakers to automatically reroute requests to the proper engine when acceptance rates drop. Crucially, you will own the cost/performance trade-off, ensuring our AWS architecture scales to millions of events without letting token costs spiral. Act as a Technical "Force Multiplier". Beyond building systems, you’ll mentor the squad through design reviews and collaborate with the Research team to ensure our agent is not just state-of-the-art, but production-safe. Experience and qualifications
Strong Backend & Cloud Expertise. You have 8+ years of experience building scalable resilient systems. You are strong in AWS architecture and know how to design for high availability and low latency with cost in mind. Polyglot Ecosystem (Java & Python). You are comfortable operating in a hybrid environment: Java for deep integration with the Core Sonar analysis engine and Python for the AI/Agent reasoning layer. You should be strong in one and capable in the other. We make use of various LLM tools, and are looking for people not afraid of different languages. Systems Thinking. You understand that "AI" in production is mostly "Systems Engineering." You know how to handle asynchronous event-driven architectures, manage queues, and design sandboxed execution environments to run untrusted code safely. AI & LLM Integration. You have practical experience building applications that orchestrate LLMs (OpenAI, Anthropic, or hosted open models). You understand the backend complexity of RAG, context window management, and prompt engineering at scale. Data-Driven Mindset. You care about the "Why." You want to build telemetry pipelines that distinguish between "Bad UX" and "Bad Code," focusing on the metrics that matter. Additional comments
This role is based in Geneva. We are unable to consider candidates unwilling to be in Geneva, but we are willing to relocate the right candidate.
We value diversity, equity, and inclusion
At Sonar, we believe that our diversity is our strength. We are a global company that values and respects different backgrounds, perspectives, and cultures. We are committed to fostering a diverse and inclusive work environment where everyone feels valued and empowered to contribute their best. We are proud to be an equal opportunity employer and welcome all qualified applicants, regardless of race, color, religion, gender, gender identity or expression, sexual orientation, national origin, genetics, disability, age, or veteran status.
If you need any accommodation, please reach out to us at .
All offers of employment at Sonar are contingent upon the results of a comprehensive background check and reference verification conducted before the start date.
We do not currently support visa candidates in the US.
Applications that are submitted through agencies or third party recruiters will not be considered.
