Outcomes & KPIs Key Outcome(s) Owned
Sub-hundred-millisecond SLM and detection service execution speeds across high-volume tenant workloads.
Highly resilient, auto-scaling GPU inference cluster infrastructure with automated CI/CD and regression gating.
Success Metrics / KPIs
P95 and P99 detection latency for prompt injection (PI) and classification services.
GPU pod readiness time and cluster auto-scaling efficiency under burst traffic.
Service availability and rate-limiting guardrail enforcement per tenant.
Automated MLOps/LLMops pipeline throughput and deployment turnaround time.
What You'll Do
Architect, build, and optimize production serving pipelines for Small Language Models (SLMs) and traditional ML models using model quantization, batching, and GPU acceleration.
Own and scale backend microservices (e.g., sds-service-prompt-injection, sds-service-ddr) delivering real-time threat detection.
Build consumption-based monitoring, tenant rate-limiting, and runtime guardrail infrastructure.
Manage and optimize GPU cluster replicas (Kubernetes / Pod infrastructure) to minimize cold-start times and manage infrastructure costs.
Build continuous evaluation, automated retraining pipelines, model CI/CD, and automated regression gates to test model detection quality against live benchmarks.
Refactor legacy service boundaries, service-to-service auth flows, and backend database configurations.
Requirements
5+ years of software engineering experience building distributed backend systems or high-throughput microservices.
Proven track record in model serving and inference optimization (e.g., TensorRT, vLLM, ONNX Runtime, Triton Inference Server).
Strong programming experience in Python, Go, or C++.
Hands-on expertise with Kubernetes, container orchestration, GPU scheduling, and infrastructure auto-scaling.
Solid understanding of CI/CD pipelines, MLOps/LLMops principles, and cloud infrastructure (AWS/GCP).
Preferred / Nice to Have
Experience building consumption-based rate limiting or multi-tenant guardrail architectures.
Familiarity with database refactoring, auth protocols, and backend technical debt remediation.
Background in cybersecurity, threat detection engines, or DLP/PII filtering platforms.
Experience benchmarking runtime vs. accuracy trade-offs between local SLMs and cloud LLMs (e.g., Vertex AI).
Experience with AWS, GCP, Terraform, Kubernetes, and Linux.
Follow us on LinkedIn https://www.linkedin.com/company/menlo-security/!
Why Menlo?
At Menlo, we don't settle for the status quo — in our technology or our culture. How we think and act is just as important as what we build. Our culture is defined by five core mindsets: Proactive Leadership, Straight Talk, United Impact, Elevated Talent, and Customer-Compelled. We take ownership and drive outcomes without waiting to be told. We communicate directly and seek hard truths. We break down silos and win together. We hold a high bar for ourselves and the people around us. And we treat every customer interaction as mission-critical. If you're someone who sees it, owns it, solves it, and does it — you'll thrive here.
All qualified applicants will receive consideration for employment without regard to race, sex, color, religion, sexual orientation, gender identity, national origin, protected veteran status, or on the basis of disability.
TO ALL AGENCIES: Please, no phone calls or emails to any employee of Menlo Security outside of the Talent organization. Menlo Security’s policy is to only accept resumes from agencies via Ashby (ATS). Agencies must have a valid services agreement executed and must have been assigned by the Talent team to a specific requisition. Any resume submitted outside of this process will be deemed the sole property of Menlo Security. In the event a candidate submitted outside of this policy is hired, no fee or payment will be paid.
