Role Summary
Sophos Firewall is the flagship product of Sophos, operating in the network security domain to protect customer network traffic when deployed in router or switch mode.
As part of the NSG Engineering Group, you will join the Firewall Assistant team in a senior capacity. This team focuses on leveraging AI to interpret natural language queries, deliver intelligent recommendations, and automate firewall operations while upholding enterprise-grade security and compliance standards. The team’s vision is to evolve firewall management from a complex, error-prone manual process into an intuitive, conversational experience—enabling security teams to operate with greater speed, accuracy, and confidence.
What you will do
- Own and deliver the production AI harness (assistant-style features: agent loop, tools, guardrails, and evaluations) while driving developer automation and large-scale migration work.
- Take ownership of end-to-end systems and not prompt-only or notebook-only work.
What you will bring
- 7-8 years of software engineering experience
- Hands-on experience with at least two of the following languages: Python, Go, or Java, including the ability to read and modify code beyond your primary language when needed.
- Experience building and deploying production-grade agents using frameworks such as LangGraph, Claude Agent SDK, OpenAI Agents, or Pydantic-AI, including tool integrations, MCP servers (e.g., FastMCP), and multi-agent orchestration. Familiarity with Temporal, Pydantic, FastAPI, and related ecosystems is highly valued.
- Proven experience designing and owning end-to-end distributed systems, including APIs, data models, storage strategies, caching, queuing, and handling failure scenarios and scalability challenges.
- Strong foundation in computer science fundamentals such as data structures, concurrency, networking, and databases, with the ability to reason clearly about latency, throughput, and consistency trade-offs.
- Ability to write clean, testable, and maintainable code, with a focus on robust API design, backward compatibility, and operational excellence.
- Demonstrated experience developing evaluation frameworks, tracing mechanisms, and regression safeguards for LLM-based systems.
- Experience building evaluation frameworks, tracing systems, and regression safeguards for LLM-based applications, beyond reliance on vendor-provided tools.
- Familiarity with prompt caching, KV-cache optimization, model routing, and context engineering to improve cost efficiency and system reliability.
- Experience fine-tuning models (e.g., LoRA, QLoRA, SFT, DPO) for real-world use cases, with measurable improvements in quality—considered a strong differentiator.
- Exposure to deploying solutions across on-premises, cloud, and hybrid environments, with an understanding of constraints, networking, identity, and upgrading cycles.
- Hands-on experience with AWS and large-scale cloud deployments, including infrastructure as code (IaC), networking, identity management, observability, cost/performance optimization, and safe rollout practices.
- Understanding of high-volume telemetry and alerting patterns, along with safety mechanisms for agent-driven actions such as approvals, dry runs, blast-radius control, and audit trails.