Senior AI Engineer

General Employee; San Francisco (CA); On-site; USD 140,000 - 230,000; Full time

Job summary

General Employee builds AI agents for service businesses. We are seeking a senior engineer to own the production systems behind AI employees working in live customer workflows. You will make real-time agents reliable enough to answer customers, take action, recover from failure, and represent a business every day.

Our stack includes LiveKit, OpenAI Realtime, Twilio, SIP, Supabase, Next.js services, and integrations with calendars, CRMs, and operating systems customers already use.

Qualifications

• Experience owning production LLM, agent, real-time, voice, or similarly failure-sensitive systems.
• Strong TypeScript, backend, API, data-modeling, and distributed-systems judgment.
• Rigorous approach to evaluation, observability, incident diagnosis, and production verification.
• Comfort moving between model behavior, application code, provider infrastructure, and customer workflow details.
• Care with privacy, security, and the consequences of an agent acting on behalf of a real business.

Responsibilities

• Own the production systems behind AI agents, ensuring reliability in real-time customer interactions.
• Harden latency, interruption handling, timeouts, retries, and recovery paths.
• Build integrations for booking, lead capture, messaging, telephony, CRM, and practice-management workflows.
• Create evaluations, simulations, scenario suites, and provider smokes that test behavior experienced by customers.
• Inspect calls, transcripts, logs, and persisted state to diagnose runtime behavior.
• Improve observability, security boundaries, data handling, and failure visibility across services.
• Collaborate with the founder and customers to translate operating rules into durable agent behavior.

Skills

Production systems

TypeScript

Distributed systems

Observability

Privacy & security

LiveKit familiarity

Job description

General Employee builds AI employees for service businesses. Our product answers calls, texts, and emails, books appointments, follows up with leads, and works inside the software a business already uses.

We are a small, founder-led team shipping into live customer workflows where reliability, judgment, and measurable outcomes matter more than impressive demos.

About the role

We are looking for a senior engineer to own the production systems behind AI employees. This is not a model-demo role. You will make real-time agents reliable enough to answer customers, take action, recover from failure, and represent a business every day.

Our current stack combines LiveKit, OpenAI Realtime, Twilio and SIP telephony, Supabase, Next.js services, and integrations with the calendars, CRMs, and operating systems our customers already use.

What you'll own

• Design and ship real-time voice agents, including instructions, tool contracts, workflow state, handoffs, memory, and retrieval.
• Harden turn-taking, interruption handling, latency, noise behavior, transfers, retries, timeouts, and recovery paths.
• Build trustworthy integrations for booking, lead capture, messaging, telephony, CRM, and practice-management workflows.
• Create evaluations, simulations, scenario suites, and provider smokes that test the same behavior customers experience.
• Inspect calls, transcripts, logs, and persisted state to diagnose actual runtime behavior rather than guessing from prompts.
• Improve observability, security boundaries, data handling, and failure visibility across agent and application services.
• Work directly with the founder and customers to turn messy operating rules into durable agent behavior.

What success looks like

• Agents complete customer work predictably across normal, ambiguous, and failure scenarios.
• Failures are visible, diagnosable, and recoverable instead of being hidden behind plausible conversation.
• Customers can trust the system with more responsibility because its behavior is proven by evidence.

What we're looking for

• Experience owning production LLM, agent, real-time, voice, or similarly failure-sensitive systems.
• Strong TypeScript, backend, API, data-modeling, and distributed-systems judgment.
• A rigorous approach to evaluation, observability, incident diagnosis, and production verification.
• Comfort moving between model behavior, application code, provider infrastructure, and customer workflow details.
• Care with privacy, security, and the consequences of an agent acting on behalf of a real business.
• Familiarity with LiveKit, WebRTC, SIP, OpenAI Realtime, or telephony is useful, but strong adjacent experience is welcome.

How we work

AI is part of our daily workflow. We use coding, design, and research agents to move quickly, but we own the judgment, verification, and production outcome. You will work directly with the founder and customers, have broad autonomy, and carry work from an ambiguous problem to a deployed result.

Equal opportunity

General Employee is an equal opportunity employer. We evaluate candidates on evidence of their work, judgment, and potential, and we welcome applicants from all backgrounds.

If you need a reasonable accommodation during the application process, email [e-mail].