Principal Machine Learning Engineer - CX
Job Description
Cisco’s CX AI Incubation and Agents team builds the foundational AI platform that powers intelligent customer experiences across Cisco, including Cisco IQ and customer deployments. This hybrid role in San Jose, CA includes 3 days per week on site, with a focus on model and inference layers, evaluation for shipped AI behavior, agentic and orchestration infrastructure, and production AI services and APIs.
What you’ll do
As a Principal Machine Learning Engineer, you will help design, build, and scale multi-tenant AI services that can operate across cloud, on-premises, and air-gapped environments. You will translate new capabilities into production architectures tuned for scalability, latency, memory, and compute across GPU, CPU, and edge systems.
- Design and build robust multi-tenant AI services with scalable APIs, versioning, backward compatibility, and quota management.
- Select and evaluate inference runtimes, orchestration frameworks, and serving technologies.
- Optimize production architectures for scalability, latency, memory, and compute across GPU, CPU, and edge environments.
- Implement cost optimization by improving token efficiency, model selection, inference performance, and compute utilization.
- Define measurable engineering standards that drive cost efficiency.
- Lead architecture and design reviews, resolve complex technical trade-offs, mentor senior engineers, and influence platform direction through hands-on execution.
What you bring
- PhD + 7 years of related experience, or Master’s + 11 years, or Bachelor’s + 14 years.
- Experience designing and building production-grade, multi-tenant distributed systems in Go, Python, or similar, including high throughput service APIs, asynchronous or event-driven architectures, observability, and reliability engineering at scale.
- Experience architecting and deploying production LLM/SLM inference runtimes, including continuous batching, KV-cache management, or speculative decoding.
- Experience architecting and deploying production RAG systems and agentic workflows using vector databases.
- Experience building automated evaluation frameworks for AI systems, such as golden test datasets, LLM-as-a-judge approaches, drift detection, and CI/CD regression gates.
Tools and technologies
Go, Python, LLM, SLM, vector databases, CI/CD, vLLM, SGLang, Triton, LangGraph.
Compensation and benefits
The starting salary range posted for this position is $296,300.00 to $374,800.00 (projected salary range for new hires in U.S. and/or Canada locations, not including incentive compensation, equity, or benefits). The hybrid role is located in San Jose, CA.
Cisco also provides benefits including medical, dental and vision insurance (U.S. employees subject to plan eligibility rules), a 401(k) plan with Cisco matching, paid parental leave, and short and long-term disability coverage. Additional benefits include basic life insurance, grants of Cisco restricted stock units that vest following continued employment for defined periods, 10 paid holidays per full calendar year, 1 floating holiday for non-exempt employees, and 1 paid day off for an employee’s birthday.
Time off and leave benefits also include paid year-end holiday shutdown, 4 paid days off for personal wellness determined by Cisco, vacation and sick time provisions, and up to 10 optional paid volunteer days per full calendar year. For non-sales roles, employees may be eligible for annual bonuses subject to Cisco’s policies, and sales plan employees may earn performance-based incentive pay on top of base salary subject to the applicable Cisco plan.
Application details
The application window is expected to close on 12/17/2026.
Preferred qualifications
- Core contributions to open-source inference, serving, or agent-orchestration projects such as vLLM, SGLang, Triton, LangGraph.
- Hands-on experience optimizing inference workloads for resource-constrained compute environments including CPU, NPU, small GPU footprints, and on-prem hardware appliances.