Performance Engineer, Hippocratic AI

• 10+ years in performance engineering with a solid software development and SRE background
• Proven track record building automated performance test harnesses (Locust, k6, Gatling, JMeter, or custom tooling)
• Deep hands-on experience with PostgreSQL performance tuning (execution plans, indexing strategies, connection pooling, autovacuum) and Redis (eviction policies, clustering, pipelining)
• Solid grasp of distributed systems fundamentals: queueing theory, tail latency, backpressure, cascading failures
• Fluency with observability tools: Prometheus, Grafana, Cloudwatch, etc
• VoIP / Telephony
• Working knowledge of SIP, RTP/RTCP, and RTCP-XR
• Hands-on experience with VoIP testing tools (SIPp or equivalent) and interpreting call quality reports
• Ability to diagnose audio degradation across network, codec, and application layers
• Familiarity with MOS scoring methodologies (PESQ, POLQA, or E-model)

Desirable

• We're hiring a Performance Engineer to own performance across our entire stack
• You'll build the automated harnesses that keep us honest, continuously measuring every model, microservice, and infrastructure component, and be our expert voice on VoIP quality characterization
• This is a high-impact, highly technical role that cuts across ML infrastructure, backend services, and real-time communications
• Build Automated Performance Harnesses
• Design and maintain automated performance testing frameworks covering the full system: LLM inference, REST/gRPC microservices, and infrastructure components including PostgreSQL, Redis, message queues, and object storage
• Integrate performance suites into CI/CD so every deploy is gated against latency and throughput regressions
• Define SLIs/SLOs and build dashboards (e.g., Grafana) that give engineering teams real-time visibility into system health
• Characterize VoIP Performance
• Measure and track VoIP quality metrics: MOS scores, jitter, packet loss, latency, echo, and codec fidelity
• Build synthetic call load testing infrastructure to stress telephony paths at scale
• Correlate audio degradation signals with underlying infrastructure metrics to root-cause issues
• Drive a Performance Culture
• Partner with ML, Speech, Backend, and Infra teams to turn performance findings into prioritized engineering work
• Contribute to incident reviews where latency or audio quality was a factor
• Write runbooks, share learnings, and help other engineers instrument their own services