Senior Software Engineer - Agents

About this role

Senior Backend Engineer - Agents Location: Remote (US) or Hybrid (Boston) ABOUT US Nebulock builds hunt-first, behavior-first security operations. Enterprise security teams use us to find attackers already inside their environment by spotting behavior that's abnormal for that organization, rather than by matching signatures of attacks someone has already seen. Attackers rotate tooling constantly and it costs them almost nothing; changing how they behave is much harder. Series A, $34M total raised led by FirstMark, with Bain Capital Ventures, Decibel, Zetta, Step Function, and In-Q-Tel. Founding team out of CrowdStrike, Palo Alto Networks, and Arctic Wolf. The platform is a closed loop: - A per-customer behavioral system of record that learns what normal looks like in a specific environment and gets more useful the longer it runs - Detections built from that context, re-baselining as environments shift, maintained by a detection engineering team rather than shipped and forgotten - Autonomous hunting agents that work from a hypothesis instead of an alert, and write the detection when they establish root cause - Detection engineering agents that take the context of threat hunts run to build, test, validate and deploy behavioral detections for Nebulock customers. POSITION OVERVIEW As a Senior Backend Engineer - Agents at Nebulock, you will work on the agent infrastructure and platforms upon which our agents are built, as well as the agents that power our product. You’ll co-own our TRACE Graph https://nebulock.io/blog/trace-context-graph-behavioral-system-of-record and other knowledge stores, ensuring our agents continuously adapt to feedback and learn. This role is ideal for engineers who are relentless about iterating and improving upon non-deterministic, agentic systems. This role will collaborate closely with our detection engineers and threat hunters to ensure that we are encoding their domain expertise into our agents. EXPECTED IMPACT - Continuous iteration, evaluation, and cross-team collaboration to ensure that our agents are behaving like the expert detection engineers and threat hunters on staff at Nebulock - Design, build, and maintain scalable data structures that are optimized for our agents (our TRACE graph, embeddings, semantic search etc.) - Ensure that our agents are behaving safely and as expected - Build and maintain automated evaluations and benchmarks for our agents to ensure that we can iterate quickly while not having regressions - Provide feedback and guidance to other engineers who are working on agents at Nebulock - Step in to work on additional backend systems as needed (we are a startup and flexibility is key to our success) QUALIFICATIONS - 4+ years software engineering experience with 1+ years building customer-facing agentic systems in production - Strong backend software engineering fundamentals - Proficiency in Python, Java, Go, Rust, or similar - Strong understanding of agentic architectures and their tradeoffs - Experience designing, building, and evaluating production RAG systems - Strong experience with agent evals and agent observability - Experience with either AWS or GCP - Ability to adapt, iterate, and ship quickly - Hunger for growth and the desire to work in a low-ego environment - Able to stay up to date with the newest advancements in the LLM and agent space while maintaining a relentless focus on delivering customer value NICE TO HAVES - Cybersecurity experience - Startup experience - Experience with graph databases - Experience evaluating the performances of different LLMs COMPENSATION We offer a competitive compensation package comprised of a base salary, equity and comprehensive benefits. The base salary for this role is $160,000 - $200,000. Compensation packages are determined based on a number of factors including but not limited to, role related experience, skillset, and geographic location.   BENEFITS - Health, dental, and vision coverage - Unlimited PTO - Remote working flexibility - Life Insurance - 401k

View the original posting ↗