SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
RapidSOS is the leading public safety AI company that unlocks mission-critical intelligence for emergency response. The company harnesses artificial and human intelligence to fuse life-saving data from the world's largest safety network, which includes over 700M+ connected devices and 200+ global technology companies. This intelligence is delivered to 23,000+ federal, state, and local agencies, serving one million first responders. RapidSOS has supported over one billion emergencies in 16 countries.
The Voice AI team is building the next generation of intelligent safety technology. The team builds AI agents that triage incoming alerts in seconds, relieve the burden on emergency telecommunicators by handling non-emergency calls, and provide real-time translation between callers and call-takers.
As a Software Engineer II, AI, you will own core Voice AI agent capabilities end to end—from model behavior and backend architecture to evaluation tooling. You will be responsible for end-to-end product delivery, owning Voice AI agent capabilities from problem discovery and rapid experimentation through architecture, implementation, testing, deployment, and measurement of impact across the real-time agent, APIs, web application, and cloud infrastructure.
Key responsibilities include improving the caller experience—including latency, turn-taking, barge-in, speech recognition, multilingual behavior, protocol adherence, and transfers—using production evidence and reproducible experiments. You will design fail-safe behaviors, strengthen service resilience and observability, and respond to production issues through contributing factors analysis, mitigation, and prevention. You will build and use automated evaluations, call-replay tools, production metrics, and controlled traffic trials to identify regressions, quantify trade-offs, and improve model and system behavior.
You will work across the full stack, from Python backend and AI model behavior to the React/TypeScript UI that customers and operators use to configure and monitor their AI agents. You will use AI tools daily to move faster on code, testing, call analysis, and incident triage, while keeping full ownership of architecture, safety boundaries, and production decisions.
Every improvement shipped shaves seconds off a response, and in an emergency, seconds are what save lives.
REQUIREMENTS:
- 2-3+ years of back-end or full-stack engineering experience shipping production software end to end
- Hands-on experience building AI/LLM applications
- Solid fundamentals in distributed systems, APIs, and operational reliability
- Real-time voice, streaming media, or conversational agent experience (strong plus)
- Strong Python proficiency, including async programming
- Flexibility to work full-stack, including TypeScript/React frontend (prior experience a plus or open to learning)
- Ability to evaluate non-deterministic AI behavior using datasets, automated metrics, qualitative call review, controlled experiments, and clear statistical reasoning
- Bias for action balanced with care: self-starter who moves fast with ownership mindset, states intent rather than asks for permission
- Product-minded problem solving: able to turn ambiguous problems into simple solutions, make evidence-based trade-offs
- Clear written and verbal communication with collaborative streak
- Fluent with AI-assisted coding, always looking for new ways to use AI to work faster
- Willingness to join weekly on-call rotation and travel 1-2 times per quarter for in-person team gatherings (mostly within Europe)
