ID 4524 – Senior AI Engineer – Fullstack, GenAI & Voice Agents

The company is hiring a Senior AI Engineer to design, build and ship production-grade full-stack and GenAI applications for our clients, with voice agents as a key specialization.
You will own the end-to-end lifecycle of full-stack and AI features: backend services and web interfaces, LLM orchestration and retrieval, and real-time voice agents (telephony, audio streaming, speech recognition and synthesis), along with evaluation and deployment. You will work closely with clients, product owners and domain experts, and set technical direction for less experienced engineers.

Key responsibilities:
● Build the full-stack application end to end: backend APIs, data persistence, and web interfaces (chat and call dashboards, transcripts, agent configuration screens).
● Design and build GenAI applications such as RAG assistants and agentic workflows, and lead voice agent delivery: real-time audio streaming, speech-to-text, LLM-driven dialogue, text-to-speech, and telephony or WebRTC integration.
● Implement agentic workflows (tool calling, multi-step planning, function orchestration) that let agents take actions and retrieve knowledge (RAG), with frameworks such as Semantic Kernel, LangGraph, Autogen or Pipecat.
● Build with voice and speech platforms and APIs such as Azure AI Speech, Azure
● AI Foundry voice models, OpenAI Realtime API, Azure Communication Services, Twilio, LiveKit, Deepgram and ElevenLabs.
● Engineer natural conversations: low-latency streaming, voice activity detection, barge-in and interruption handling, turn-taking, endpointing, and graceful fallback and escalation to a human.
● Own evaluation: build offline and online eval harnesses for LLM and voice quality (end-to-end latency, word error rate, task completion, groundedness, cost), golden datasets and regression gates in CI.
● Take systems to production: latency and cost optimization, observability and call tracing, rate limiting, and failover across speech and model providers.
● Handle sensitive data correctly: PII redaction in text, audio and transcripts, call-recording consent and retention, compliance-aligned controls, secrets management and RBAC.
● Partner with clients, product managers and domain experts to turn ambiguous requirements into shippable features.
● Mentor engineers, review designs and code, and set standards for prompt engineering, evaluation and LLMOps across teams.
● Contribute to client-facing technical discussions, architecture reviews and estimates.

Required qualifications:
● 8+ years of software engineering experience, with 3+ years of hands-on GenAI/LLM work, including systems you designed, shipped and supported in production (not prototypes or internal demos).
● Strong backend development in both Python and .NET (C#); comfortable with FastAPI, ASP.NET Core or similar service frameworks, async programming and containerization (Docker, AKS or Azure Container Apps).
● Full-stack capability: front end in a modern framework (React, Angular or Vue) with TypeScript, REST APIs and background jobs, authentication and authorization (OAuth 2.0, RBAC), relational or document data stores, and cloud deployment.
● Experience with at least one orchestration framework: Semantic Kernel, LangChain/LangGraph, LlamaIndex, Autogen or a voice agent framework such as Pipecat.
● Hands-on RAG or knowledge-retrieval experience: embeddings, vector databases (Azure AI Search, pgvector, Pinecone, Qdrant or similar), hybrid retrieval and citation/grounding.
● Demonstrated ability to evaluate LLM and voice systems quantitatively (latency, word error rate, task completion, groundedness) and improve them iteratively.
● Solid grounding in software fundamentals: API design, data modeling, testing, CI/CD (Azure DevOps or GitHub Actions) and infrastructure as code (Bicep or Terraform).
● Hands-on experience building AI voice agents: speech-to-text, text-to-speech, real-time audio streaming and LLM orchestration, ideally with at least one shipped phone or web voice product.
● Practical experience with at least one speech or voice platform, such as Azure AI Speech, Azure AI Foundry voice models, OpenAI Realtime API, Deepgram, ElevenLabs, Twilio, LiveKit, Vapi, Retell or similar.
● Real-time communication experience with WebSockets and/or WebRTC.
● Experience working with sensitive data and security controls, including PII handling in text, audio and transcripts.
● Clear written and verbal communication; able to explain trade-offs to non-technical stakeholders.

Preferred qualifications:
● Experience with Azure AI Speech, Azure Communication Services, Azure AI Foundry (including its voice models), Azure OpenAI Service or Azure AI Search; Azure certifications such as AI-102 or AZ-305.
● Speech-to-speech and realtime multimodal models; fine-tuning or customizing speech models (custom vocabulary, custom neural voice); multilingual and accent-robust voice solutions.
● SIP/telephony and contact-center, IVR or CCaaS integrations (Genesys, Five9, Amazon Connect, Twilio Flex), including latency, jitter and codec handling; call analytics such as summarization, sentiment and QA scoring.
● Experience applying voice or LLM solutions to workflows in a regulated industry such as healthcare, financial services or insurance.
● Familiarity with Model Context Protocol (MCP) and tool ecosystems for agents.
● Multimodal or document-heavy RAG (images, tables, scanned documents) using Azure Document Intelligence.
● Prior experience in a consulting or client-delivery setting.

What we offer:
● Ownership of GenAI, voice and full-stack solutions that ship to real end users, not internal demos.
● A small, senior team with direct access to leadership and clients.
● Modern cloud and AI stack with budget for experimentation on new language and voice models and services.