Senior Full Stack Engineer (Conversational AI and Voice Applications)

The Senior Full Stack Engineer will lead the technical assessment and development of an AI-powered voice practice application for a US-based training business serving insurance agencies. The solution will simulate realistic customer conversations, allowing learners to practice handling objections and responding naturally in real time. You will own the architecture and delivery of a two-way voice experience, from browser audio and conversational AI to backend services, scoring, and coaching reports. Working across frontend, backend, and AI integrations, you will turn approved training content into configurable scenarios and reliable assessment workflows. This is a hands-on, project-based engagement with an initial paid assessment followed by phased development and validation. The role is fully remote within India, with scheduled collaboration during US business hours.

• Assess existing LMS capabilities, reusable prototypes, and potential platform extensions before recommending a configured or custom solution.

• Document feasibility, dependencies, delivery estimates, technical risks, and operating-cost assumptions.

• Define the application architecture and implementation approach based on business needs and available resources.

• Build a functional two-way voice prototype supporting responsive audio, natural turn-taking, interruptions, session state, and recovery from common failures.

• Develop an AI customer persona with approved facts, objections, and objectives, ensuring consistent behavior across different learner responses.

• Design natural conversation endings without revealing expected answers or guiding learners toward predetermined responses.

• Integrate speech and language model services using a real-time speech-to-speech approach or a combination of speech recognition, LLM, and speech synthesis.

• Convert approved training materials into configurable, versioned scenarios and assessment rubrics, keeping the customer simulation separate from post-session evaluation.

• Investigate integration options with Absorb LMS, including learner identity, browser microphone permissions, and application launch workflows.

• Validate embedded access and a separate-tab fallback, then implement the agreed authentication and completion approach.

• Ensure the application works across agreed browsers and LMS launch paths.

• Implement evidence-based scoring and balanced coaching using approved weights, exclusions, and serious-error rules.

• Calculate scores programmatically and ensure generated coaching PDFs accurately reflect the underlying assessment results.

• Store learner, agency, scenario, rubric, and attempt records in a structured database.

• Generate branded PDF reports and agreed data exports while preserving information for future reporting integrations.

• Distinguish actual audio-based measurements from estimates derived from text, including optional speaking metrics only when validated and agreed.

• Implement agency and learner usage allowances, reset rules, alerts, and limits.

• Handle concurrent requests, retries, and failed attempts without duplicate charges or assessment results.

• Protect agency data boundaries, credentials, and recordings through appropriate access controls, retention, and deletion policies.

• Maintain useful application logs while avoiding unnecessary exposure of personal information.

• Deploy and maintain the application, manage environments and secrets, monitor latency and usage, and resolve technical defects.

• Conduct automated testing and debugging across browser, backend, and AI services.

• Demonstrate progress weekly, maintain an agreed task tracker, and proactively communicate blockers and technical trade-offs.

• Deliver source code, configuration, deployment instructions, monitoring guidance, cost assumptions, and operational documentation.

• Support user testing, scoring calibration with content specialists, defect resolution, and final handover.

The ideal candidate is a hands-on full stack engineer with proven experience delivering conversational or voice-based applications, strong frontend and backend development skills, and the ability to integrate AI services into reliable production-ready products.

• Typically 5–8 years of software engineering experience, with relevant delivery experience taking priority over tenure.

• Proven ownership of at least one working two-way voice application or conversational product involving spoken interaction. You should be able to explain your contribution, architecture, latency considerations, interruption handling, failure recovery, and usage costs.

• Strong frontend development skills with TypeScript and React or a comparable framework.

• Practical experience with browser audio capture and playback, asynchronous UI behavior, microphone permissions, and session error handling.

• Backend development experience using Python (such as FastAPI) or Node.js with TypeScript.

• Familiarity with REST APIs, WebSockets and/or WebRTC, authentication, background processing, and relational data modeling.

• Practical experience integrating LLM and speech services through real-time speech-to-speech or a combination of speech recognition, language models, and speech synthesis.

• Strong prompt and context design skills, including structured model outputs, scenario grounding, and evaluation against approved facts.

• Ability to recognize when an AI system should acknowledge missing information rather than generate unsupported responses.

• Experience with third-party API integrations, secure session management, SQL databases such as PostgreSQL, PDF generation, and reliable data export workflows.

• Familiarity with automated testing, debugging, Git, code reviews, deployment pipelines, Docker or equivalent packaging, and at least one cloud environment.

• Clear written English and the ability to translate business requirements into small, testable releases while explaining assumptions and trade-offs to non-technical stakeholders.

• A degree in a relevant discipline or equivalent practical experience.

• Near-term availability is preferred.

• Absorb LMS or another learning management system.

• LTI 1.3, OIDC/SSO, or secure external-tool launch integrations.

• Multi-tenant SaaS architecture and agency-level data separation.

• Rubric-based AI evaluation, training, or coaching products.

• Insurance workflows and document retrieval solutions.

• Integrations with Google Sheets, BigQuery, HubSpot, or Zapier.

• Avatar integration, if introduced in a later project phase.

You do not need experience with every platform listed; strong expertise in one coherent technical stack is sufficient.