Software Engineer, Voice AI / Backend
At a glance
Mid-level Backend Engineering role at Clera. Berlin · full-time · €85,000–€160,000 base.
techroles summary, based on the employer's posting.
What you'll do
- Build production voice AI with retrieval-augmented generation, model routing and checks of transcription providers before release
- Run a distributed call scheduler that handles simultaneous calls, time zones, retries, race conditions and provider limits
- Extend agents across voice, messaging and email while keeping memory and state consistent between live and asynchronous conversations
- Develop systems that retain customer context, and tools for configuring agents, investigating problems and connecting business software
- Strengthen tracing and recovery across calls, workers and providers so failures can be found and addressed
What you bring
- You have shipped and owned distributed systems in production, with solid software fundamentals and end-to-end project delivery
- You have production experience with Python, FastAPI or TypeScript, alongside data stores such as PostgreSQL or Redis
- You have worked with real-time systems, audio or voice infrastructure, plus concurrency, load balancing or fault tolerance
- You have put LLMs, AI models or agent-based systems into production and used AWS, GCP or another listed cloud
Who this fits
This suits a mid-level engineer who can take projects from product choices through architecture and production, with a focus on customer and business outcomes. The role is on-site in Berlin, Germany. Visa sponsorship and relocation support are available.
From the employer
About the Role
Join a small, fast-moving engineering team building voice AI agents for customer-facing workflows. You will own projects end to end, from product decisions and technical architecture to reliable production systems.
What You'll Do
Build reliable voice AI systems, including retrieval-augmented generation, dynamic language model routing, and testing transcription providers before rollout.
Design and operate a distributed call scheduler that manages concurrent calls, time zones, retries, race conditions, and provider limits.
Extend agents across voice, messaging, and email, with shared memory and state across real-time and asynchronous conversations.
Develop memory and reasoning systems that preserve context across customer interactions.
Build AI tools that configure agents, investigate issues, and connect with business software such as CRMs, calendars, and email.
Improve observability and resilience so failures across calls, workers, and providers can be traced and addressed.
What We're Looking For
Production experience building and owning distributed systems, with strong software fundamentals and a track record of delivering projects end to end.
Production experience with Python, FastAPI, or TypeScript, plus data stores such as PostgreSQL or Redis.
Experience with real-time systems, audio processing, or voice infrastructure, and with distributed-system challenges such as concurrency, load balancing, and fault tolerance.
Experience shipping LLMs, AI models, or agentic systems in production, and working with AWS, GCP, or Azure.
Familiarity with Temporal or other workflow orchestration tools, message queues, frontend frameworks such as React, and observability tools such as OpenTelemetry or Honeycomb is useful.
Comfort learning unfamiliar domains, taking initiative, and focusing on customer and business outcomes.
Compensation & Benefits
€85,000 to €160,000 annually, plus equity. Visa sponsorship is available.
Location
On-site in Berlin, Germany. Relocation support is available.