Mitra is building the world’s first AI phone-calling assistant — an agent that makes and answers real phone calls on your behalf, so you can be part of every conversation even when you can’t pick up. Founded in 2025 and backed by leading investors, Mitra’s mission is to amplify human presence — to let people scale themselves across real-world interactions, relationships, and moments that matter.
We’re looking for a software engineer to design and scale the core infrastructure that powers Mitra’s real-time agent runtime – the system that enables AI to converse naturally over live phone calls, manage memory and personalization, and orchestrate tools across the telephony network.
Responsibilities
- Architect and implement Mitra’s agent runtime for phone-based interactions — orchestration, state management, and multi-turn memory for personalized calls.
- Build high-reliability real-time systems for audio streaming, STT/TTS pipelines, and multi-agent coordination at scale.
- Design sandboxed and monitored execution for external integrations (Twilio, LiveKit, ElevenLabs, Supabase, Azure).
- Develop evaluation frameworks for latency, call quality, and conversation success metrics; automate regressions and A/B tests.
- Help bring new agent capabilities from prototype to production.
Qualifications
- 5+ years of experience building large-scale distributed or real-time systems.
- Strong coding skills in TypeScript, Python, or Go, with experience designing APIs and service infrastructure.
- Experience with audio streaming, WebRTC, or telephony APIs (Twilio, SIP, PSTN).
- Familiarity with containers, concurrency, and observability in production.
- Deep curiosity and ownership across layers – from backend systems to voice UX.
Nice to Have
- Experience with agent orchestration pipelines (task routing, planning, long-term memory, or contextual embeddings).
- Exposure to voice synthesis, speech recognition, or LLM-driven conversation engines.
- Background in reinforcement learning from human feedback or automated evaluation for conversational quality.