About this role
Practice by Numbers seeks an AI Software Engineer to build its AI Receptionist, a patient-facing conversational product spanning voice, SMS and web chat for dental practices. This hands-on individual contributor owns features from development through production, including the backend services, APIs and integrations behind them. The role reports to the Lead Engineer — AI.
Responsibilities
- Build and maintain Python backend services and RESTful APIs using FastAPI or Django. Design PostgreSQL schemas and efficient queries; use Redis for caching and session state. Implement asynchronous and event-driven workflows involving queues, webhooks, WebSockets and background workers.
- Develop LLM-driven conversation flows with tool calling, multi-turn state and context handling. Refine prompts, implement RAG for practice-specific questions, and build patient-facing guardrails that prevent medical advice and unverified data disclosure. Evaluate changes and regression-test AI quality while balancing response quality, latency and cost.
- Integrate with Dentrix, Open Dental, Eaglesoft and internal APIs; implement secure authentication, including OTP-based patient verification. Follow HIPAA-compliant data handling, logging and storage practices.
- Own logging, metrics, alerting, production debugging, and unit and integration tests. Work with Product Management on requirements and edge cases; participate in planning, reviews and documentation while collaborating with US-based stakeholders.
Required qualifications
2–6 years of professional software development experience, including production backend services and APIs and improving systems after release. Practical experience with LLM applications, prompt design, tool calling and handling model outputs is required; substantial personal or open-source work counts, but tutorial-level work does not. Strong Python, PostgreSQL, async Python, API and integration experience; working knowledge of AWS or GCP/Azure; and familiarity with CI/CD, LLM evaluation, RAG and agentic patterns are expected. Candidates should communicate clearly and work independently amid evolving requirements.
Preferred
Conversational AI, voice or telephony, speech APIs, LLM orchestration frameworks, healthcare/HIPAA knowledge, observability tools, task queues, multi-tenant SaaS or B2B products, and open-source contributions.
This full-time role is in-office in Gurugram only. Hours are primarily 10 AM–5 PM IST, with occasional evening overlap with US teams until 9–11 PM IST as needed.