Architect Performance

Team Overview

The Oracle Cloud Infrastructure (OCI) team is revolutionizing end-to-end data center

planning, design, construction and operations by developing innovative software

solutions. We aim to create a seamless bridge between digital and physical

infrastructure, ensuring efficient management and automation. By standardizing data

center building and implementing advanced analytics, we streamline data center

design and construction, enabling rapid scalability. With a focus on automation and AI

integration, OCI is poised to lead the industry in cost-effective, efficient data center

management, making it accessible and adaptable for future technological

advancements.

Role Overview

We are seeking a technically exceptional Performance Architect to lead, mentor, and

enforce high standards in system performance engineering. As a hands-on technical

leader, you will set methodology, drive diagnosis and solutions for performance

bottlenecks at every layer (kernel, network, storage, runtime), and evangelize best

practices across a highly visible, cross-organizational scope.

Key Responsibilities

• Deep Linux internals: scheduling, run queues & preemption, context switches,

virtual memory (THP/hugetlbfs, page cache, reclaim/compaction), MMU/TLB

shootdowns, softirqs/IRQs.

• Networking expertise: TCP/UDP/QUIC, congestion control (e.g., CUBIC/BBR),

socket options, NIC offloads (TSO/GSO/GRO/LRO), RSS/RPS/RFS, NAPI/busy poll, routing/VLAN/VXLAN basics.

• Coding & code reading: strong C/C++, low-level concurrency, lock/contention

analysis, cache-aware data structures, microbenchmarks; quickly read

unfamiliar C and propose measurable improvements.

• Linux tuning at scale: sysctls, cgroups v2, CPU/IRQ affinity, NUMA placement,

huge-page policies, qdisc selection (fq/fq_codel), net/IO queue tuning.

• Observability & instrumentation: low-overhead tracing; metrics/exemplars;

flame-graph culture; Prometheus/OpenTelemetry/Grafana; actionable SLO aligned alerts.

• Databases: query plans, indexing, buffer/cache behavior, lock/wait analysis,

read/write amplification; OS/IO/network interactions.

• Reliability via performance controls: admission control, backpressure, circuit breaking, rate limiting; SLO/SLA thinking and tail-tolerance.
• Impact communication: crisp RCAs with before/after metrics, clear tuning

guides/playbooks; ability to influence and coach across teams.

Basic Requirements

• 15+ years of professional experience in software development, with strong

expertise in system design, distributed systems, and automation platforms.

• 10+ years of experience specifically building and delivering software solutions

for data center environments, including DCMI, BMS, and data center

automation.

• Proven track record of architecting large-scale, mission-critical systems with

high availability and performance.

• Strong knowledge of modern programming languages (e.g., Java, C++, Go,

Python) and cloud-native technologies (containers, Kubernetes, service mesh).

• Hands-on experience with APIs, integration frameworks, and microservices

architectures.

• Deep understanding of monitoring, telemetry, and control systems used in data

centers.

• Strong communication skills with the ability to influence stakeholders and

present to executives.

• Experience working in global, cross-functional organizations with complex data

center environments.

Preferred Requirements

• Knowledge of infrastructure-as-code, DevOps pipelines, and automated

deployment strategies.

• Familiarity with large-scale cloud provider environments.
• Advanced degree in Computer Science, Electrical Engineering, or related field

Key Responsibilities

• Deep Linux internals: scheduling, run queues & preemption, context switches,

virtual memory (THP/hugetlbfs, page cache, reclaim/compaction), MMU/TLB

shootdowns, softirqs/IRQs.

• Networking expertise: TCP/UDP/QUIC, congestion control (e.g., CUBIC/BBR),

socket options, NIC offloads (TSO/GSO/GRO/LRO), RSS/RPS/RFS, NAPI/busy poll, routing/VLAN/VXLAN basics.

• Coding & code reading: strong C/C++, low-level concurrency, lock/contention

analysis, cache-aware data structures, microbenchmarks; quickly read

unfamiliar C and propose measurable improvements.

• Linux tuning at scale: sysctls, cgroups v2, CPU/IRQ affinity, NUMA placement,

huge-page policies, qdisc selection (fq/fq_codel), net/IO queue tuning.

• Observability & instrumentation: low-overhead tracing; metrics/exemplars;

flame-graph culture; Prometheus/OpenTelemetry/Grafana; actionable SLO aligned alerts.

• Databases: query plans, indexing, buffer/cache behavior, lock/wait analysis,

read/write amplification; OS/IO/network interactions.

• Reliability via performance controls: admission control, backpressure, circuit breaking, rate limiting; SLO/SLA thinking and tail-tolerance.
• Impact communication: crisp RCAs with before/after metrics, clear tuning

guides/playbooks; ability to influence and coach across teams.